a... u. i .353. n ”a 3:... 1 .12. :C 1...? 5.... 5.13.... Int. ii. ‘9 . .2 r1. .4. Snub _ ‘ ‘1 2(2on This is to certify that the thesis entitled LOCATING ALTERNATIVE SAND SOURCES FOR MICHIGAN’S FOUNDRY INDUSTRY: A GEOGRAPHICAL APPROACH presented by Bradley R. Schrotenboer has been accepted towards fulfillment of the requirements for the MS. degree in Geography Major'AProféssor’H Signature 10/115 4205;? Date MSU is an Affirmative Action/Equal Opportunity Employer JR. PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5/08 K‘IProj/AccspraSJClRCIDateDue.indd LOCATING ALTERNATIVE SAND SOURCES FOR MICHIGAN ’S FOUNDRY INDUSTRY: A GEOGRAPHICAL APPROACH By Bradley R. Schrotenboer A THESIS Submitted to Michigan State University in partial fiJlfillment of the requirements for the degree of MASTER OF SCIENCE Geography 2008 ABSTRACT LOCATING ALTERNATIVE SAND SOURCES FOR MICHIGAN’S FOUNDRY INDUSTRY: A GEOGRAPHICAL APPROACH By Bradley R. Schrotenboer Numerous large coastal dune fields occur on the western coast of Lower Michigan. These dunes are an important ecological, geological, and recreational resource in the state. They also serve as a significant source of foundry sand for Michigan’s automotive industry and thus have been intensively mined. Although Michigan contains extensive sand deposits besides those in coastal dunes, no studies have yet investigated alternative foundry sources from a distinct geographical perspective. Acceptable alternative sand deposits must be sufficiently large and close to transportation networks to be economically viable. Using a GIS, water-well log stratigraphic data were employed to estimate sand thickness and rail line data were used to determine accessibility of deposits. Based on this information, 53 sites in 16 counties were selected, sampled, and tested for appropriate physical and chemical characteristics to determine their viability as inland sources of foundry sand. Results indicate that many cubic kilometers of inland sand exist in close proximity to existing transportation networks. Three regions that show the most potential to be inland sand sources for the foundry industry are: 1) Wexford and southeastern Grand Traverse Counties, 2) northern Newaygo and southern Lake Counties, and 3) central Alger County. Sand in these areas will likely require preprocessing but should nonetheless be considered as a feasible and more ethically responsible alternative to mining coastal dunes. ACKNOWLEDGEMENTS I would like to thank the Michigan State University Geography Department, Ford Motor Company, and Tanya Cabala of Alliance for the Great Lakes for providing the resources necessary for me to complete this project. I would also like to thank my advisor, Alan Arbogast, and committee members, Randy Schaetzl and Jay Harman, for their constructive comments and suggestions on the many drafts of this thesis, that in the end have made me a much better writer. Most of all, I would like to thank my wonderful wife Abbie, who provided invaluable and much needed support through all phases of this project. iii TABLE OF CONTENTS LIST OF TABLES .................................................................................. vi LIST OF FIGURES ................................................................................ viii CHAPTER 1 INTRODUCTION .................................................................................... 1 CHAPTER 2 LITERATURE REVIEW ........................................................................... 6 Lake Michigan Coastal Dunes ............................................................. 6 Foundry Sand ............................................................................... l7 Pre-Regulatory History of Sand Mining in Michigan ................................. 23 Sand Mining in Michigan During the Regulatory Era ................................. 26 Environmental Ethics, Sand Mining, and the Lake Michigan Coastal Dunes. .....35 CHAPTER 3 STUDY AREA ...................................................................................... 4O Geology and Paleoclimate ................................................................ 42 Climate ....................................................................................... 46 Vegetation ................................................................................... 46 Sofls .......................................................................................... 47 CHAPTER 4 METHODS .......................................................................................... 49 Preliminary Mapping Methods ........................................................... 49 Field Methods .............................................................................. 51 Laboratory Methods ........................................................................ 54 Mapping Sand Thickness .................................................................. 60 CHAPTER 5 RESULTS AND DISCUSSION .................................................................. 63 Economic Analysis ........................................................................ 63 Transportation Costs .............................................................. 64 Sand Thickness and Volume ..................................................... 66 Physical Analysis of Sands ............................................................... 76 Grain Shape ........................................................................ 80 Grain Size Distribution ........................................................... 90 25-micron Clay .................................................................... 99 Chemical Analysis of Sands ............................................................ 104 pH .................................................................................. 1 04 Acid Demand Value ............................................................. 108 Overall Suitability ........................................................................ 1 12 iv CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS ............................................ 123 APPENDIX A SAMPLE LOCATIONS .......................................................................... 127 APPENDIX B SAMPLE DEPTHS ................................................................................ 129 APPENDIX C OUTWASH GRAIN SHAPE, GRAIN FINENESS NUMBER, AND 25-MICRON CLAY CONTENT ................................................................................... 133 APPENDIX D SAMPLE LOCATION RANKING CRITERIA ............................................................. 145 LIST OF REFERENCES .......................................................................... 146 4.1 4.2 5.1 5.2 5.3 5.4 5.5 5.6 5.7 A] Bl C1 C2 C3 LIST OF TABLES General sieve ranges for foundry sand. Specific users’ requirements can vary based on the casting technique used and the specifications ofthe finished product56 Multipliers used to determine AF S Grain Fineness Number .......................... 59 Conventional and AF S grain shapes for inland dune samples ........................ 89 Grain Fineness Numbers for inland dune samples. Standard deviations represent variability among multiple samples from the same location. . . .98 Percentage 25-micron clay for inland dune samples ................................. 103 pH and Acid Demand Value for glacial outwash locations (deepest sample below 3.0m) ............................................................ 107 pH and Acid Demand Value for inland dune locations (deepest sample below 3.0m) ............................................................ 108 Suitability rankings for all glacial outwash locations. A maximum score of 28 would indicate all variables are within foundry parameters. The subtotal column is the sum of values for grain shape, GFN, sieves, and 25- micron clayl 15 Suitability rankings for all dune locations. A maximum score of 28 would indicate all variables are within foundry parameters. The subtotal column is the sum of values for grain shape, GFN, sieves, and 25-micron clay .............. 117 Sample locations .................................................................................................. 127 Sample depths ...................................................................................................... 129 Conventional and AF S grain shapes for glacial outwash samples. Primary roundness and sphericity are the first values listed for each sample ................... 133 Grain size distribution statistics for glacial outwash samples. Standard deviations represent variability among multiple samples from the same location ................................................................................................................ 137 Percentage 25-micron clay for glacial outwash samples ..................................... 14] vi D1 Sample location ranking criteria. Higher ranks identify locations that are more suitable to the foundry industry based on these criteria ....................... 145 vii 1.1 2.1 2.2 2.3 2.4 2.5 3.1 3.2 3.3 3.4 4.1 4.2 4.3 5.1 5.2 LIST OF FIGURES Images in this thesis are presented in color. Surficial geology of Michigan. Potential sand sources include dunes, outwash, lacustrine sediments, and alluvrum3 Coastal dune areas along the Lake Michigan shore in Michigan ............................ 7 Foundry sand core used for casting engine block components .............................. 18 Average grain size requirements of Michigan dune sand users ...................... 22 Average chemical impurities limits for foundries surveyed by Ayres and Chapman (1978) ............................................................................ 22 Annual tonnage of coastal dune sand mined in Michigan, 1978 to 2005 .............. 33 Potential non-bedrock Michigan sand sources ........................................... 40 Michigan bedrock formations composed entirely or partly of sandstone. Most Lower Peninsula formations are buried by glacial drift and not exposed at the surface ..................................................................... 41 Counties in which sand samples were collected for this study ....................... 42 Ice readvances in Michigan after the Wisconsin glacial maximum (~ 18 ka). During the readvances of 15.5 ka and 13 ka all of the Upper Peninsula was covered by glacial ice ..................................................... 44 Locations from which inland sand was collected and analyzed. The Wexford Sand Co. mine is labeled ...................................................... 52 General grain size distribution for foundry sand. Specific users’ requirements can vary based on the casting technique used and the specifications of the finished product ...................................................... 55 AF S standard comparison grain shapes used in grain shape determination .....59 All areas greater than 3.25 km (2 miles) from a Great Lakes shoreline within 8 km (5 miles) of an existing rail line ........................................... 65 Minimum estimated cumulative sand thickness between the surface and bedrock. Data are minimum estimates because not all water-wells used in the estimation process penetrate to bedrock ............................................ 68 viii 5.3 5.4 5.5 5.6 5.7a 5.7b 5.7c 5.8 5.9 5.10 5.11 5.12 5.13 5143 Estimated cumulative thickness of sand between the surface and a depth of 23 m (75 ft) .............................................................................. 70 Number of water wells, by public survey township. Public survey townships are usually 36 square miles in area, 6 miles on a side........................72 Michigan bedrock wells. Drift wells are not shown. Bedrock wells are common in the southern half of the Lower Peninsula and in the Upper Peninsula, but are virtually absent from the northern Lower Peninsula. . . . . . .73 Lowerrnost stratigraphic unit of Michigan water-wells. Locations in pink are wells completed in sand, while those in gray are wells completed in bedrock or non-sand drlft . 75 Sample locations in southern and central Lower Michigan ........................... 77 Sample locations in northern Lower Michigan ......................................... 78 Sample locations in the Upper Peninsula ................................................ 79 Conventional geologic and engineering grain shape classification matrix with sample results plotted. Ideal grain shape for foundry use is subangular ...... 81 AF S grain shape results. The predominant grain shape was subangular in over 95% of the samples ................................................................... 82 Conventional geologic and engineering grain shape classification matrix for glacial outwash samples. Ideal grain shape for foundry use is subangular ..... 83 Outwash sand grains from Crawford County (Crawford 4, sample C). Predominant grain shape was largely subangular, but grains larger than approximately 400 microns in diameter often showed distinct rounding. Arrows identify more rounded grains. Magnification is 40x.... 84 Sand from Wexford Sand Company’s inland sand mine, 25x magnification. Notice the increased roundness of the larger grains but the predominance of subangular shape ........................................................................... 85 Sand sample from a depth of 6 m (sample D) at Wexford 1. Arrows point to compound sand grains. Magnification: 25x ........................ 86 Sand from an inland dune site in Arenac County (Arenac 2, sample D). Magnification is 25x ....................................................................... 88 ix 5.14b 5.15 5.16 5.17 5.18 5.19 5.20 5.21 5.22 5.23 5.24 5.25 5.26a 5.26b Glacial outwash sand in Crawford County (Crawford 2, sample D). Although average grain size is slightly larger than at the inland dune site in Arenac County (Arenac 2), grain shape is virtually indistinguishable. Magnification is 25x ....................................................................... 88 Conventional geologic and engineering grain shape classification matrix for inland dune samples .................................................................................... 90 Grain Fineness Number (GF N) results. Target range for foundry sand is 47 to 53. Labeled locations have one or more samples within the target range ............................................................................................................ 92 Frequency distribution of grain fineness number for glacial outwash samples ..... 94 Variability in grain size distribution based on the standard deviation of GFN for sites with samples from multiple depths. Low standard deviations are desired, as they signify more consistency of GFN at a given location. . ....94 Results of the sieve analysis. Labeled locations are acceptable on 7 or more sreves96 Clay content results for all sample locations. Labeled locations are those in which all samples had less than 1.0% claleO Frequency distribution of 25-micron clay content for glacial outwash samples ................................................................................................................. 100 25-micron clay content of the deepest sample from each location. . . . . .............101 pH of the deepest sample taken at each location (if below 3.0 m) ................. 105 Frequency distribution of Acid Demand Value (ADV). Ideal values are below 10.00 ................................................................................. 109 Acid demand value (ADV) for all tested locations. Locations with values less than 10.00 are best suited for foundry sand and have been labeled.............111 Primary and secondary suitability zones in northwest Lower Michigan. Boundaries are generalized based on railroad corridors and estimated sand thickness ............................................................................. 118 Primary and secondary suitability zones in the Upper Peninsula. Boundaries are generalized based on railroad corridors and estimated sand thickness ............................................................................. 1 19 Chapter 1 INTRODUCTION Michigan’s coastal dunes have long been an important ecological, geological, and recreational symbol in the state. The best developed of these dunes occur primarily along the eastern and northern shores of Lake Michigan (Michigan Legislature, 2001). They ~ harbor a biological community that is found nowhere else in the state, and very few other places around the world (Harman and Arbogast, 2004; Michigan Legislature, 2001; Michigan Department of Environmental Quality, 1979). The fragility of these diverse ecosystems makes them especially susceptible to human disturbance (van Dijk and Vink, 2005; Michigan Department of Environmental Quality, 1979). Geologically, these dune complexes preserve important clues to the post-glacial development of Michigan’s coastal landscape. While theoretically renewable on a geologic timescale, they are essentially nonrenewable within human timeframes (Harman and Arbogast, 2004; Arbogast et al., 2002b). In addition, these dunes are the centerpiece of coastal recreation in the state. They draw visitors from far away, and are one of the most highly used and highly visible landscapes in Michigan (van Dijk and Vink, 2005; Michigan Legislature, 2001; Michigan Department of Environmental Quality, 1979). Michigan’s coastal dunes are also heavily used by the state’s industrial sand mining industry, particularly to supply mold and core sand used in iron foundries. Most of the sand mined in Michigan for foundry use comes fiom these dunes (Lake Michigan Federation, 1999). Coastal dunes are ideal sources of sand for use in foundries for a variety of reasons. They have a small range of grain-sizes, being particularly well-sorted, and have very few chemical and physical impurities (Michigan Legislature, 2001; Lewis, 1975). Variability of grain shape and grain size is fairly low across deposits. In terms of volume, they are large enough to be profitable for mining (Ayres et al., 1978; Kelly, 1971). In addition, they are generally accessible to both rail and barge transport, the two preferred methods of transporting industrial sand (Marrone, 2005, pers. com.). A collision of competing values, however, has occurred with Michigan’s coastal dunes. The industrial value of the coastal dunes to sand mining companies is in direct conflict with the environmental and recreational value of these dune areas to citizens. Many environmental groups strongly support protecting the coastal dunes from sand mining, including Alliance for the Great Lakes, Annis Water Resources Institute, Michigan Environmental Council, Michigan United Conservation Clubs, Preserve the Dunes, Inc., and Western Michigan Environmental Council. Coastal dune sand, however, is also extremely important to Michigan’s industrial economy, specifically the auto industry. The conflict occurs because, as currently practiced, industrial sand mining diminishes the recreational and ecological value of coastal dune areas (Michigan Department of Environmental Quality, 1979). Herein is found the tension of this issue: supporting either side without providing an alternative for the other results in a loss of value (tangible and/or intangible) to the people of Michigan. By presenting a viable alternative to mining coastal dune sand, I suggest that the recreational and ecological value of Michigan’s coastal dunes can be preserved. At the same time, a new source of industrial value to Michigan’s economy would be identified, to replace coastal dunes. Sand deposits in Michigan are not restricted to coastal dunes. In fact, coastal dune sand covers only a very small portion of the state. Other sources of sand include inland dunes, glacial outwash plains (sandurs), beach ridges (and former beach ridges), sandy lacustrine sediments, and alluvium (Figure 1.1)--deposits formed by a diverse range of geomorphic processes. In addition to surficially visible sand deposits in Michigan, subsurface sand deposits can be identified using data fi'om the hundreds of thousands of water-well stratigraphic records collected by the Michigan Department of Environmental Quality (MDEQ). 92'W 90°W 88'W 86°W 84'W 82'W 0 50 100 Kilometers l—l—l 46°N - - Dune sand - Glacial outwash . - Ice-contact glacial outwa - I - Lacustrine clay and silt - Lacustrine sand and gravel “.N h - Peat and muck - Alluvium Glacial till - Bedrock - Water 42'N *- 42°N 90°W 88'W 86'W 84'W 82°W Figure 1.1: Surficial geology of Michigan. Potential sand sources include dunes, outwash, lacustrine sediments, and alluvium (modified from Farrand and Bell, 1982 and Dorr and Eschman, 1970). Environmental groups, industry groups, and the state government have all expressed a need to study whether Michigan’s inland sand resources have the potential to replace coastal dune sands in the foundry process (Johnson, 2005, pers. comm; Lake Michigan Federation, 1999; Michigan Department of Natural Resources, 1984). To date very few studies have been conducted on this subject. The most recent published study to examine suitable altemative sources was conducted by Sundeen (1978a, b) three decades ago. No studies have been from a distinctly geographical perspective. New technologies, such as Geographic Information Systems and computerized geostatistical modeling, have been developed that can analyze large amounts of data in ways not previously possible. Michigan’s most recent statewide surficial geology map (Farrand and Bell, 1982) and the NflDEQ collection of water-well log stratigraphic records were both compiled since the last study on this topic. The purpose of this study is to determine whether Michigan’s widespread inland sand resources have the potential to be used by the foundry industry. Such potential is based on economic, physical, and chemical criteria of the sand and sand deposits, including grain size, grain shape, clay content, pH, acid demand value, deposit volume, and deposit accessibility. These general characteristics are the most important in determining whether sand is viable for foundries. This research uses these characteristics as a means of comparing untapped inland sand deposits to “ideal” foundry sand, and to an existing inland sand mine that has been shown to be viable for foundry use. By using this methodology a direct comparison to dune sand was unnecessary. The results of this study provide valuable data for environmental groups and industry groups alike. Not only do the results provide an overview of certain physical and chemical characteristics of Michigan’s inland sand, they detail the spatial patterns and accessibility of these sand resources. The results add to previous research on the use of inland sand and provide a platform for more detailed analyses and feasibility studies of specific deposits. They also direct future research to those parts of the state estimated to have the most and best sand. Chapter 2 LITERATURE REVIEW This review provides the relevant background on the current status of foundry sand usage in Michigan. It begins with the geomorphic history of Lake Michigan coastal sand dunes, which are the current source of almost all foundry sand mined in the state. Following this discussion is a description of foundry sand characteristics/spedfications and how they correlate with Michigan sand dunes. Next, the framework includes the history of sand mining in Michigan in both the pre-regulatory and post-regulatory era. Concluding the review is an overview of the emerging environmental ethics movement and its effects on the sand mining debate in Michigan. Lake Michigan Coastal Dunes Coastal dunes occur in every county bordering the eastern shoreline of Lake Michigan, as well as in Schoolcrafi and Mackinac Counties along the lake’s northern shore (Figure 2.1). These dunes have been studied for over 100 years. The earliest studies of the dunes focused on their ecology. In 1899, Cowles published the first dune study of the region, specifically on the dunes of northwest Indiana and southwest Michigan (Figure 2.1). He noted that dunes in this region usually overlie either the beach or bluffs of clay or gravel. Cowles (1899) also argued that dune formation appeared to be regulated by sand supply and the relation of the coastline to the prevailing wind direction. Although he made many qualitative descriptions of the dunes, his primary focus was on plant response to rapidly changing dune environments. 88°W 86°W 84°W 46°N Ma istique 46°N \ f U 44." _ Ludington 44." Montague ran Haven Holl nd 42°N L I - 42°N Nati Tolesto 0 40 30 Kilometers $ I——i—I 88°W 86°W 84°W Figure 2.1: Coastal dune areas along the Lake Michigan shore in Michigan (developed from Arbogast, et al., 2004, Farrand and Bell, 1982, and Lewis, 1975). The number of studies on Lake Michigan coastal dunes increased significantly in the 19305 and 19405. Stevenson (1931) studied dunes near Manistique, Michigan (Figure 2.1) and was the first to observe a possible correlation between episodes of coastal dune formation and lake levels. She specifically compared the low foredune ridges adjacent to the lake and higher parabolic dunes farther inland. Stevenson (1931) concluded that the foredunes formed when lake levels fell. She also noted that, while the massive parabolic dunes appeared deformed and distorted in comparison to the foredune ridges, both dune forms were generally parallel to the modern shoreline. From this relationship, she theorized that the parabolic dunes began as foredune ridges that were eroded by wave action. This erosion caused blowouts and the distortion of the linear ridges into collections of large parabolic dunes. Following Stevenson (1931), Scott and Dow (193 7) used location, direction, and soils to determine relative ages of parabolic dunes, foredunes, and beach ridges near Upper and Lower Herring Lakes (Figure 2.1). Using their landscape positions they estimated that the large parabolic dunes north of the lakes were formed during the Nipissing high stand of Lake Michigan. Smaller foredunes later formed on the post- Nipissing bay-mouth bar that now isolates Lower Herring Lake from Lake Michigan (Scott and Dow, 193 7). The following year, Dow (193 8) studied dunes that mantle high morainal bluffs along the northeast shore of Lake Michigan. He referred to these dunes as “perched dunes” because of this geomorphic relationship. Unlike Scott and Dow (1937), Dow ( 1938) did not focus on their ages, but on the processes by which they formed. He examined the perched dunes most closely at Sleeping Bear Point (Figure 2.1) and hypothesized four possible origins for the eolian sand at that locality: 1) sand of beach origin that accumulates by wind action, 2) sand of beach origin accumulating due to wind action and shore recession, 3) sand derived from the moraine itself, and 4) sand derived from both the moraine and the beach. By looking at the effects of wave action, wind action, and gravity on sand composition at the bottom, mid-section, and crests of the bluffs, Dow (193 8) determined that the largest portion of perched dune sand accumulates following erosion of the bluff’ 5 upper crest due to wind or undercutting by wave action. Larger particles are dislodged and fall downslope by gravity while sand-sized particles are entrained by the wind and carried over the bluff crest to an area of accumulation. Erosion is greatly increased during periods of high lake level, when waves directly out into the base of the bluff. A much smaller contribution to the overall sand accumulation is the migration by wind action of sand grains up the steep bluff over a long period of time. In the 19405, a number of dune studies investigated both Lake Michigan dune processes and origins. Calver (1940) studied the roundness of Lake Michigan dune sands, developing an exponential function to describe the rate of rounding as grains move fi'om beach to dune. Hi5 conclusion, however, was that differences in rounding between beach sand and dune sand were too minor to be used as an accurate means of differentiation. In the same year, Dow (1940) classified ventifacts associated with perched dunes at Sleeping Bear Point (Figure 2.1). In his presidential address to the Michigan Academy of Science, Arts, and Letters, Scott (1942) discussed the complex problems involved in explaining the origin of Lake Michigan’s coastal dunes. These included the origin of the sand that makes up the dunes, the numerous dune forms present, and the effects of differential isostatic rebound. Two Ph.D. dissertations published in 1946 attempted to explain the origin of dune and dune-related landforms along Lake Michigan. Tague (1946) studied the Grand Marais embayment, in Berrien County (Figure 2.1), and mapped a succession of shoreline features and dunes associated with the Glenwood, Calumet, Algonquin, Nipissing, and Algoma phases of Lake Michigan. Calver (1946) studied the Platte and Crystal Lake depressions, along the northeast shore of Lake Michigan (Figure 2.1). He identified alternating periods of coastal landform erosion (Algonquin and Algoma phases of Lake Michigan) and of construction (Nipissing and present phases). Erosional phases were associated with the presence of wave-cut bluffs and few dunes, while constructional phases were periods of dune formation. In 1958, Olson (1958a, b, c) published 3 significant studies on dune growth and movement in Michigan. His first study (Olson 19583) described the effects of vegetation on wind velocity on Michigan coastal dunes. He showed that the presence of dune vegetation, such as marrarn grass (Ammophila breviligulata), can reduce mean wind velocity near the plant to zero up to a height of 1 cm. This forms a zone surrounding the dune vegetation in which wind velocity is below the threshold required for sand transport. Sand deposition is thus initiated and dune formation begins. Olson’s (1958b) second study showed that dune vegetation can be used to determine the recent geomorphic history of a dune. Marrarn grass (Ammophila breviligulata) initiates sand deposition after colonizing barren sand surfaces and exhibits seasonal growth cycles in heavily depositional, sandy environments. Buried leaf sheath 10 bases on the marram grass stems mark the positions of previous summers’ dune surfaces. Olson (1958b) also showed that certain vegetation communities are related to different phases in dune development. For example, some communities become established only after active deposition of sand has ceased. Thus, the ages and types of shrubs and trees on a dune provide information about its geomorphic state. In his third study, Olson (19580), like Stevenson (1931), theorized that lake level played a prominent role in dune building along Lake Michigan. Olson (1958c) compared a sequence of beach ridges to a radiocarbon chronology of the Great Lakes to determine the relationship between lake level and beach ridge formation in northwest Indiana. He concluded that beach ridges form when a drop in lake level is paired with sufficient sand deposition. Over the past 2,700 years this combination of events has produced beach ridges about once every 30 years. These beach ridges were preserved by the growth of stabilizing vegetation (Olson, 1958b, c). The vegetation trapped eolian sand, causing the beach ridges to become the cores of new beach-margin dunes. In their book Geology of Michigan Dorr and Eschman (1970) summarized the work of Olson (1958a, b, c) and compared the histories of inland dunes to the large coastal dunes. They supported Olson’s (1958c) view that lake level plays a prominent role in coastal dune growth and estimated the age of Michigan’s large coastal dunes at 4,500 years, corresponding to formation during the Nipissing Great Lakes stage. Inland dunes, however, are older than coastal dunes and thus related to earlier stages of Michigan’s glacial lakes, especially Glacial Lake Saginaw and Glacial Lake Chicago. Regardless of their differing ages, Dorr and Eschman (1970) argued that coastal and inland parabolic dunes in Michigan have formed in a similar fashion. 11 The next major study on Michigan’s coastal dunes was conducted by Buckler (1979), who developed a dune classification system based on a variety of characteristics (e.g., morphology, relief) and how they relate to the underlying strata. According to Buckler (1979), the most extensive period of dune formation along Lake Michigan occurred as lake levels were dropping from the Nipissing high stand approximately 4,000 years ago, to the level of Lake A1goma(~ 3,000 yrs. BP). This hypothesis was consistent with previous work by Stevenson (1931) and Dorr and Eschman (1970). During the 1990s a new wave of Lake Michigan coastal dune research began that initially focused on the evolution of beach ridges and associated foredunes along Lake Michigan. The first major study of this period was conducted by Thompson (1992), who investigated the extensive beach ridge complex at the Toleston Beach along the southern shore of Lake Michigan (Figure 2.1). Thompson (1992) measured the elevation of foreshore deposits underlying beach ridges as proxy indicators of lake level. In addition, he determined radiocarbon ages of basal peats in the inter-ridge swales. With these two data sets, he was able recreate lake level curves and estimate rates of beach ridge formation and shoreline progradation. Results showed an overall decreasing trend in lake level from the end of the Nipissing II phase of Lake Michigan (4200 cal. yrs. BP, 180.5m msl) to the present. However, this overall decreasing trend was interrupted by quasi- periodic lake-level high-stands that varied on three distinct scales: 1) a ~30 year variation of 0.5 to 0.6m, 2) a ~150 year variation of 0.8 to 0.9m, and 3) a ~ 600 year variation of 1.8 to 3.7m. Beach ridges along the northern shore of Lake Michigan were studied in a similar fashion by Thompson and Baedke (1995) and Delcourt et a1. (1996). Thompson and 12 Baedke (1995) showed that beach ridge development in this area was generally similar to that along the southern shore over the past 2,600 years. Differences were restricted to the ~ 600 year lake level variation, which caused more coastal erosion around the northern lake margin. The beach ridge sequence is thus more complete along the southern part of the lake. In a similar study, Delcourt et a1. (1996) employed tree-ring dating on trees growing on ridges closest to the modern shoreline, and radiocarbon dating of sediment from inter-ridge swales farther inland to compare the time periods of beach ridge formation with historic climate data and paleoclimate models. Using these techniques they estimated that ridges have formed approximately every 72 years during the last 5,400 years. This cycle correlates closely with the 65 to 70 year Northern Hemisphere climate oscillation identified by Schlesinger and Ramankutty (1994). Because beach ridge formation occurs during periods of falling lake levels (Olson, 1958c; Thompson, 1992; Thompson and Baedke, 1995) and falling lake levels generally correlate with rising temperatures (Fraser et al., 1990), Delcourt et a1. (1996) presented these ridges as climatic indicators. Much recent dune research has focused on the geomorphic histories of the large parabolic dune complexes that line much of the Lake Michigan shore. The first such study was conducted by Arbogast and Loope (1999), who investigated four parabolic dune complexes between Grand Haven and Ludington (Figure 2.1). Unlike perched dunes (e.g. Dow, 1938), which overlie morainal drift up to 150 m above current lake level, these parabolic dunes overlie lacustrine sediments only 15-25 m above the present- day lake. It had generally been assumed (e.g. Buckler, 1979; Dorr and Eschman, 1970) 13 that they formed during the Nipissing high stand of Lake Michigan. In order to test this hypothesis, Arbogast and Loope (1999) radiocarbon dated charcoal fiom buried soils in the uppermost lacustrine or lowermost eolian sediments at each location. These dates provided maximum-limiting estimates for the beginning of dune development. Results from one site (Nordhouse Dunes, near Ludington; Figure 2.1) suggested that dune formation began during the Nipissing high stand of ancestral Lake Michigan, between 4,820 and 4,410 cal. yrs. BP. Results from the other three locations, however, suggested dune development began afier the Nipissing high stand. Arbogast and Loope (1999) concluded that dunes at all four of these locations likely started forming during and immediately after periods of high lake level, perhaps after the low lake-terrace bluffs were undercut by wave action. This finding is in line with Dow’s (1938) model of perched dune development and not consistent with Olsen’s (195 8c) foredune model. Next, Loope and Arbogast (2000) dated buried soils to test Thompson and Baedke’s (1995) hypothesis that ~150 year cycles in lake level high stands initiated dune- building along eastern Lake Michigan. In addition, they aimed to reconcile Olsen’s ( 195 8c) foredune model with Anderton and Loope’s (1995) perched dune model of dune development for Lake Michigan dunes. Radiocarbon dates were obtained from 75 buried soils exposed at 32 dune locations overlying moraines, outwash plains, and lake terraces along the eastern shore of Lake Michigan. These dates were then used to approximate when dune reactivation occurred after soil-forming periods. Comparisons between the time periods of dune reactivation and the lake level curve of Lake Michigan (from Thompson and Baedke, 1999) indicate that peaks in dune- building periods along the eastern shore of Lake Michigan are closely associated with 14 ~150-year cycles of high stands proposed by Thompson and Baedke (1995). Dune- building cycles were similar for dunes on morainal headlands, outwash plains, and truncated lake terraces. These results support the application of Anderton and Loope’s (1995) model of maximum dune-building during periods of high lake levels to this region. Olson’s ( 1958c) foredune model (in which sand is supplied to dunes primarily during periods of low lake levels) is restricted to the southern and southwestern shores of the lake. Loope and Arbogast’s (2000) radiocarbon dates also revealed that the majority of dune-building along northeastern Lake Michigan occurred in the late Holocene. In an ~ 50 km stretch of shoreline near Sleeping Bear Point (Figure 2.1), over 75% of the sand volume was above buried soils less than 1,500 cal. years old. This finding contradicts earlier hypotheses that most dune-building occurred during or immediately after the Nipissing high stand (Buckler, 1979; Dorr and Eschman, 1970; Stevenson, 1931). Following the work of Arbogast and Loope (1999) and Loope and Arbogast (2000), Van Oort et al. (2001) and Arbogast et al. (2002b) reconstructed the geomorphic history of large coastal dunes at Van Buren State Park and near Holland, respectively (Figure 2.1). Excellent exposures existed at each of these localities in which numerous buried soils were contained. Radiocarbon dates showed that large-scale deposition of eolian sand began shortly after the Nipissing stage of Lake Michigan at both of these sites. Van Oort et al., (2001) and Arbogast et al. (2002b) determined that most eolian sand had been deposited at both Van Buren and Holland between 4,000 and 2,500 cal. years BP. This period of dune growth was punctuated by brief (~ ISO-year) periods of stability in which Entisols formed. This period of rapid dune growth was followed by a 15 major pause in sand supply that lasted from about 2,000 cal. years BP to ~ 500 cal. years BP. During this period of time a well-developed soil with Spodosol characteristics developed. Dune reactivation has since added an additional 10 meters of eolian sand and raised the total dune heights to ~ 40 meters. This well-developed soil was studied extensively by Arbogast et al. (2004) and was informally labeled the Holland Paleosol. Geographically, the Holland Paleosol was found to intermittently extend from the Indiana Dunes area north to Montague, Michigan (Figure 2.1), and perhaps even farther north. Multiple radiocarbon dates have put the age of this buried soil between 3,000 cal. years BP and 300 cal. years BP. The reason for this period of stabilization is not yet known, but one hypothesis is that sand supply to this section of Lake Michigan dunes was diminished during this period because the active shorezone was farther west. The movement of the active shorezone would have buffered the dunes and allowed them to stabilize (Arbogast et al., 2004). Such a hypothesis is consistent with findings by Thompson (1992) that beaches were rapidly prograding along the southern shore of Lake Michigan during this time interval. The reconstructed geomorphic histories of these dunes show that dune building has occurred over many thousands of years in response to complex variables (Arbogast et al., 2004; Arbogast et al., 2002; Van Oort et al., 2001; Loope and Arbogast, 2000; Arbogast and Loope, 1999). Some variables, such as lake level, are relatively constant around the entire Lake Michigan basin, while others, such as sand supply, can vary significantly over relatively short distances. Current models of formation suggest that foredune ridges form relatively frequently (Baedke and Thompson, 1992), but massive parabolic dunes result from episodic sand accumulation over thousands of years 16 (Arbogast et al., 2004; Arbogast et al., 2002b; Van Oort et al., 2001). The results have been large accumulations of a valuable natural resource in areas that are easily accessible. It is clear from their geomorphic histories that these large parabolic dunes cannot be considered renewable resources on human timescales. The combination of many conditions (high lake level, sufficient sand supply, strong onshore winds, etc.) that led to their formation cannot be expected to occur with high frequency (Arbogast et al., 2004; Arbogast et al., 2002b; Van Oort et al., 2001). Even given the right conditions, the dunes took thousands of years to reach their present stature (Arbogast et al., 2004). And while beach ridges and associated foredunes can form on a decadal scale (Baedke and Thompson, 1992), they do not provide sufficient volume to replace the larger parabolic dunes as sources of foundry sand and, in any case, are usually part of the same coastal dune ecosystem. Foundry Sand Foundry sand is a subset of industrial sand that is used to create molds and cores from which metal can be cast (Figure 2.2). Such sand has been mined and used in Michigan for over 100 years (Ries and Rosen, 1908). The criteria for foundry sand can vary considerably based on the type of casting performed, as well as the particular casting process employed by a given foundry (Brown, 1936). Given that most sand casting in Michigan is iron casting only sand used in that process is considered here. In the first major study of foundry sand in Michigan, Ries and Rosen (1908) listed five major characteristics used in describing foundry sand: 1) cohesiveness, 2) refractoriness, 3) texture, 4) porosity and permeability, and 5) durability. For each of 17 Figure 2.2: Foundry sand core used for casting engine block components. these characteristics they listed ideal conditions of the sand for foundry usage. The sand mixture had to be cohesive enough to remain molded when pressed into its form. It had to display a degree of refractoriness sufficient to keep the sand grains from fusing together under the heat of molten iron. The texture had to be appropriate for the type of mold being made (a finer texture being better for smaller, smoother casts and a coarser texture being better for larger casts that need not be as smooth). Changes in texture also could affect other characteristics of the sand, such as cohesiveness and permeability. Porosity and permeability had to be sufficient to allow expanding gases to escape as the casted metal cooled. Durability of the sand was important economically as it allowed for the recycle and reuse of foundry sand (Ries and Rosen, 1908). At the turn of the 20th century no standard laboratory tests existed for measuring these characteristics of foundry sand (Ries and Rosen, 1908). By 1912, the American Foundrymen’s Association (AF A; later the American Foundrymen’s Society, and now the American Foundry Society) started assembling laboratory procedures for use in standardized testing of foundry sand (American Foundrymen’s Association, 1912). Official organized research on testing molding sand properties did not begin until 1921 (American F oundrymen’s Association, 1944). Before this time, ad hoc procedures had been developed but were often inaccurate and influenced by the persons performing the tests. In 1924, a tentative set of standards for testing moisture content, strength, permeability, fineness, and chemical properties of foundry sand was developed (American F oundrymen’s Association, 1924, 1944). In 1930, Moldenke described the ideal molding sand as being uniform-sized rounded grains of silica sand that were coated with a very thin layer of highly refractory clay (Moldenke, 1930). Revisions and updates to initial AF A testing procedures were made in 1928, 1931, and 1938. By 1944, permanent standardized procedures had been developed to test against published standards (American Foundrymen’s Association, 1944). At this time, the primary AP A tests for foundry sand were moisture, permeability, strength of bonding, fineness, and sintering point (refractoriness). Soon after, however, the natural bonding strength of foundry sand began to decline in importance, as the use of synthetic sands increased (Hofstetter, 1948). Synthetic sands are naturally unbonded sands mixed later with a bonding agent (kaolinite, bentonite, and eventually synthetic resins) The increased use of naturally unbonded sand by the foundry industry made Michigan sand, especially dune sand, increasingly desirable due to its abundance in the state (see e.g. Brown, 1936). Hofstetter (1948) listed refractoriness, proper grain size, and sufficient durability as the main requirements for foundry sand. The important characteristics of foundry sand remained largely unchanged in subsequent publications. A report by the Research Council of Alberta (McLaws, 1971) 19 listed refractoriness, bond strength, permeability, grain fineness, and moisture as the most important characteristics of foundry sand. The refractoriness of sand is influenced by its mineral constituents as well as the grain size. For example, quartz will fuse at 1710 0C, while feldspar fuses between 1200 °C and 1300 °C and magnetite fuses around 1600 °C (McLaws, 1971). A decrease in grain size contributes to an increase in surface area, making the sand more susceptible to high temperatures. Bond strength is dependent on the nature of the bonding material. Since most foundry sand is now mined naturally unbonded, grain shape has increasing importance in determining the bond strength of unprocessed sand. Angular grains provide a stronger bond and round grains make a weaker bond. According to McLaws (1971), the best permeability comes from sand with rounded and uniform grains. In Michigan, however, where the main mineral type for iron foundry sand is silica, such a grain composition is not ideal. Silica grains undergo a sudden change in volume at 573 °C. If all sand grains in a mold are similar in shape and size, they are likely to reach this temperature at the same time. This can cause a mass expansion and “scabbing” of the mold (Parkes, 1950). To prevent this from occurring, a mix of grain sizes is used so that the temperature increase is not uniform. The last two important characteristics listed in the Research Council of Alberta Report (McLaws, 1971) are grain fineness and moisture. In order to standardize the measuring of grain sizes in foundry sand, the American Foundry Society (AF S) developed a measure called Grain Fineness Number (GFN). This number represents the approximate sieve size (US. Standard Sieve) through which a sand sample would just pass if all grains were the weighted average grain size of the sample. The desired GFN 20 for foundry sand can vary based on the needs of the foundry. Moisture content must be sufficient to give the sand necessary plasticity and adhesiveness without being so much as to produce large volumes of steam during the casting process, which can lead to defects in the mold. In 1978, the Michigan Department of Natural Resources conducted a survey of Michigan dune sand users to determine their requirements for foundry sand grain shape, grain size, and chemical impurities (Ayres and Chapman, 1978). The results varied widely. Of the 132 responses, 14% specified they used rounded grains, 37% used subangular grains, 17% used angular grains, and 31% had no specific requirement for grain shape. Averaged results for grain size indicated an approximately normal distribution of grain sizes centered at US. Standard Sieve 70 (212 microns; Figure 2.2). The average maximum limits of certain chemical impurities in the sand, according to survey respondents, are shown in Figure 2.3. The specifications set forth by McLaws (1971), as well as the results of the 1978 Michigan DNR survey questionnaire, were used by Sundeen (1978a, b) in his study of Michigan sand deposits. Presently, the general requirements for foundry sand are well known, but to obtain specific requirements for foundry sand it is necessary to contact individual sand users. The current standard for laboratory sand testing procedures for foundry sand is set forth by the American Foundry Society in their Mold and Core Test Handbook, 3rd Edition (2004). This edition is the latest iteration of the standardized foundry sand testing procedures first assembled in 1924. 21 §§§ 25% - §§§ Percentage of Sand Retained NI 40,501701001140 200 Pan Illcrone (above) - U.8. Shnderd Sieve Size (below) Figure 2.3: Average grain size requirements of Michigan dune sand users (modified from Ayres and Chapman, 1978) iron timnium potassium manganese 3 magnesium O cobalt chromium calcium aluminum 0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50% Average Maximum Lln‘it Figure 2.4: Average chemical impurities limits for foundries surveyed by Ayree and Chapman (1978). 22 Pre-Regulatory History of Sand Mining in Michigan Although Michigan has long been known for its copper and iron deposits, the state has also become a leader in the mining of sand and sandstone resources in the past century (Heinrich, 1979). The history of sand and sandstone resource extraction can be divided into two main periods. From the 19th century until World War I production focused mainly on the sandstone products of building stones, scythe stones, and abrasive wheels. After World War I the focus shifted to glass sand, molding sand, and construction sand (Heinrich, 1979). Beginning in the mid-18005 numerous foundries opened in Michigan to support the major timber operations that began in the state at this time (Michigan Department of Environmental Quality, 2000). The technique of sand casting used by these foundries required that large amounts of sand be mined nearby or be imported. By the turn of the 20th century Michigan contained many foundries, but the use of Michigan sand in these foundries did not accelerate until after World War I. The first report on Michigan foundry sand was published in 1908 by the Geological Survey of Michigan (Ries and Rosen, 1908). According to this report, virtually all Michigan foundries, especially larger ones, imported sand from out of state. Foundry sand was being excavated at only a few localities within the state. At this time, sand used in iron casting was often a loamy sand, with naturally occurring clay. The textures varied, based on the size of the piece being cast. Small castings used finer sand and heavy castings used coarser sand (Ries and Rosen, 1908). Because artificially combining sand and clay to form a properly cohesive mixture was not well known or common, most sand used in Michigan foundries came from just a small 23 number of locations in Ohio, Illinois, Indiana, and Kentucky, where the correct texture occurred naturally (Ries and Rosen, 1908). Michigan sand that was used in local foundries came largely from small pits, often on private land. Some of the places in which these small pits were located were Lansing, Jackson, Battle Creek, Port Huron, Saginaw, Escanaba, Grand Rapids, and Marquette (Ries and Rosen, 1908). Ries and Rosen (1908) determined that most local sand used in Michigan foundries came from four geomorphic environments: 1) pockets in morainal drift, 2) glacial outwash deposits, 3) lake deposits, or 4) silts bordering rivers that are either present-day flood deposits or reworked glacial material. In addition, they observed that no large, uniform deposits of acceptable foundry sand had been found within the state. Any acceptable local deposits were usually found as small, discontiguous pockets, or lenses or thin beds just below the surface. Small amounts of coastal dune sand were also used at this time (Michigan Department of Environmental Quality, 2000). The need for industrial sand in Michigan foundries grew rapidly in the 19105 and 19205 due to the development of the automobile industry, the replacement of manual labor with machines, and industrial expansion fueled by World War I (Lewis, 1975). This expansion precipitated the second major investigation into Michigan foundry sand, by Brown (1936), which began in 1923. Brown’s (1936) research was an attempt to work off the findings of Ries and Rosen (1908) and determine whether the preparation of synthetically bonded molding sand fi'om naturally unbonded sand was feasible. Brown’s (1936) report spurred the move away from inland sand mining in Michigan and toward the mining of coastal dune sands by proclaiming: “It seems, therefore, that the only probable localities containing usable amounts of molding sand are the old lake beds in which the sorted material may have been 24 mixed later with a clay bond, or in small areas of moraines. In such deposits it is not to be expected that any appreciable thickness of molding sand of uniform quality will be found. . .Because of these conditions we may expect to find the best deposits of core sand or unbonded well sized sand in the sandy outwash plains, the lake deposits or beaches of old glacial lakes, and most important, in dune sands particularly along the eastern shore of Lake Michigan (pgs. 28-29).” Brown (1936) recognized that naturally bonded molding sands were insufficiently abundant in Michigan to supply the needs of local foundries. However, naturally unbonded sands were abundant and could provide ample supply to Michigan foundries if they could be used successfully in the casting process. These sands were most easily accessible in the coastal dunes (Brown, 1936). After much testing, Brown (193 6) concluded that the synthetic bonding of naturally unbonded sand produced molds of comparable quality to naturally bonded sand. This conclusion vastly increased the amount of potential foundry sand that was available within Michigan and made the unbonded sands of the coastal dunes much more desirable. In addition, the study suggested that uniform and well-sized deposits of foundry sand in Michigan could also be expected in glacial outwash areas (Brown, 1936). Expansion of industrial sand mining in Michigan increased rapidly after World War I and was largely concentrated in coastal dune areas. By 1958, almost 1.8 million tons of industrial sand resources (statistics include sand and sandstone) were being mined annually (Lewis, 1975). Kelly (1971) reported that dune sand was particularly valuable because it had high purity and uniform fineness. He also noted that most dune mining was occurring along the shore of Lake Michigan, but a few of the smaller inland dunes in Tuscola County were also being mined for foundry sand. By 1973, the amount of industrial sand mined in Michigan had risen to over 5.7 million tons mined by 10 25 different companies. Over 90% of sand supplied to automotive foundries of the US. and Canada came from Michigan and over 50% of foundry sand nationwide came from the state (Lewis, 1975). Seven companies mining along the Lake Michigan shoreline provided over 90% of the sand used by foundries of Michigan and the surrounding Great Lakes states (Lewis, 1975). Sand Mining in Michigan During the Regulatory Era In 1970, the first Earth Day, the creation of the Environmental Protection Agency, and the passage of the National Environmental Policy Act (NEPA) marked the unofficial beginning of the environmental preservation movement in the United States and ushered in a new era of environmental regulation (Palmer, 1997). By this time, millions of tons of sand had been removed from Michigan’s shorelines (Lewis, 1975). While this mining activity provided a significant economic boost to the state it also resulted in the demise of spectacular dunes and the permanent alteration of coastal landscapes (Lake Michigan Federation, 1999). The removal by mining of popular and visible dunes such as “Pigeon Hill” near Muskegon and “Maggie Thorpe” near Manistee became a rallying point for concerned citizens and an impetus for new research on the effects of sand mining. Michigan residents responded by forming new non-profit groups aimed at preserving dunes and other coastal ecosystems. The Lake Michigan Federation incorporated in 1971 with the goal of conserving and restoring Lake Michigan through policy, education, and local resource mobilization. It has since become the Alliance for the Great Lakes, and expanded its mission to include the entire Great Lakes basin 26 (Alliance for the Great Lakes, 2007). Other environmental groups focusing on Michigan dunes include Preserve the Dunes, Inc. and the Michigan Environmental Council. Research on sand mining in Michigan peaked in the mid to late 19705. Over the course of that decade, the number of agencies, companies, and private groups requesting information on Michigan’s mineral resources grew substantially (Lewis, 1975). This surge in interest sparked by the emerging environmental movement culminated in the passage by the Michigan legislature of the Sand Dune Protection and Management Act in 1976 (Michigan Legislature, Act No. 222, Public Acts of 1976). The purpose of this act was to protect, manage, and reclaim Michigan sand dune areas. In support of this purpose, the Michigan Department of Natural Resources (Michigan DNR) was required to conduct a comprehensive study and inventory of Michigan sand dunes. This task was broken into six parts: 1) an economic study of sand dune mining in Michigan, including where the sand is marketed, how the sand is used, and the quantity of sand in reserves, 2) a geologic study of non-dune sand areas in Michigan that may be able to replace dune sand for industrial uses, 3) an assessment of dune areas that should be protected through purchase by the state or other means, 4) the identification and designation of barrier dunes along the shoreline, 5) a study of methods for recycling dune sand as well as alternatives to dune sand, and 6) recommendations for protecting dune areas fi'om uses other than mining. The first of the six charges to the Michigan DNR was fulfilled by two separate studies. The Geological Survey Division of the Michigan DNR released Report of Investigation 20 in 1978, entitled An Economic Stuajr of Coastal Sand Dune Mining in Michigan (Ayres et al., 1978). Research associated with this study indicated that in 1976 27 Michigan was the nation’s leading producer of industrial sand and that 68% of the state’s industrial sand output came from the 10 companies mining coastal sand dunes. The extent of sand mining was quantified for the first time. Over 5,700 acres of coastal dunes had been mined at 26 sites, 14 active and 12 inactive. At this time a majority of dune sand users expected future demands to increase and placed estimates of the year of exhaustion of then-owned recoverable reserves between 2000 and 2010. A second study, Report of Investigation 21, examined the economic geology of sand and sandstone resources in Michigan (Heinrich, 1979). This study overviewed the various sand and sandstone deposits that occur in the state, looking at their major uses and the amount of reserves potentially available for future use. According to the study, Michigan dune sand was almost exclusively used as molding sand. Michigan supplied about 90 percent of the foundry sand used in the automotive industry and 40 percent of all foundry sand used for any purpose nationwide. Reserves were estimated at 20 to 30 years not accounting for the anticipated increase in production rate. The economic forecast was for an increase in foundry sand reclamation and a move toward sand from non-dune deposits. The second charge to the Michigan DNR was firlfilled by a two-phase geologic study of Michigan sand deposits (Sundeen, 1978a, b). This study investigated possible alternatives in Michigan to using dune sand in the foundry and glass-making industries. Samples were collected fiom easily accessible areas, with approximately 280 taken from inland dunes, 280 obtained from glacial outwash, and 20 extracted fi'om fiiable sandstones. Samples were collected from 2 to 5 feet below the ground surface. One sample per location was collected. Test results on physical and chemical characteristics 28 of the sand suggested that outwash sand may be a feasible replacement for dune sand in the foundry industry, as a number of locations had low acid demand, low AF S clay content, and appropriate grain fineness numbers. Inland dune sand did not appear to be a feasible replacement as most inland dune sand samples were too fine-grained. Six outwash areas were selected for additional testing in phase 2 of the project. Selection of these areas was difficult given the variability in results from phase 1 and locations not selected for additional testing were not ruled out as potential sand sources. The most promising results of phase 2 were from sand in southwest Lake County. However, all locations had advantages and disadvantages, and no definite conclusions were reached. Vertical variation in the sand units sampled was assumed but was not tested except at the six locations studied more extensively in phase 2 of the project (Sundeen, 1978b). Estimates of the sizes of the sand deposits encountered were only made for these locations. Otherwise only surface area estimates were made. Recommendations of the study included further evaluation of foundry requirements, of critical specifications for foundry sand, and of transportation requirements, as well as additional exploration and testing of inland sand. The third charge to the Michigan DNR was slated to be firlfilled with a three- phase study that began with a 1984 preliminary report entitled Action Plan Sand Dune Protection and Management Program (Michigan Department of Natural Resources, 1984). This preliminary report outlined the proposed three-phase study. As proposed, the study would have created a priority list of locations within designated sand dune areas that should be acquired by the state. However, the study was not completed. Natural 29 Resources Commission policy was amended on June 14, 1985 to state that the Michigan DNR places a high priority on acquiring land within designated sand dune areas and to provide a list of factors that the Michigan DNR would consider during the acquisition process, but did not assess actual locations for acquisition (Michigan Department of Natural Resources, 1985). Part four of the Michigan DNR’s task list was firlfilled with Buckler’s (1979) study on dune classification. This study developed a classification system for dune features found along Michigan’s Lake Michigan shoreline. Dune features identified for this study included parabolic dunes, linear dune ridges, dune terraces, dune platforms, domal dunes, complex dune fields, dune flats, marginal sand aprons, and interdune lowlands. These features were differentiated based on their form, relief, orientation, arrangement, and how they related to the underlying geologic formations and/or landforms. In order to clarify the meaning of the Sand Dune Protection and Management Act barrier dunes were identified according to their relative relief and their position as the most lakeward dune assemblage within the designated sand dune area. The fifth study required of the Michigan DNR was partially fulfilled by Sundeen’s (1978a, b) study of possible inland sand sources. Some work has also been done on the use of offshore, dredge materials as an industrial sand source, but generally by private groups (e.g. Peebles and Thorp, 2001). No formally published study has assessed methods for recycling or reusing Michigan dune sand in industrial processes although consultation with industrial sand users in the automotive industry has revealed that a significant percentage of molding sand is currently recycled (Marrone, 2005, pers. comm.) One of the main recommendations of the 1984 preliminary report Action Plan 30 Sand Dune Protection and Management Program (Michigan Department of Natural Resources, 1984) was the re-establishment of research programs that address the use of non-dune sand as a mineral resource in industrial processes. No such research programs have been restarted nor have any public studies on the use of non-dune sources of sand in industrial processes been published since the release of this report, despite an explicit recommendation to do so. The final task given to the Michigan DNR in the Sand Dune Protection and Management Act (Michigan Legislature, Act No. 222, Public Acts of 1976) was to recommend ways to protect the dunes from uses other than mining. The proposed three- phase study outlined in Action Plan Sand Dune Protection and Management Program (Michigan Department of Natural Resources, 1984) was to address this issue but the project did not pass the initial stages (Michigan Department of Natural Resources, 1984). However, a 1986 study by Planning and Zoning Center, Inc., in cooperation with the Michigan DNR (Wyckoff, 1986), adequately fulfilled this requirement. This study provided an overview of types of development then-occurring in designated dune areas. Non-mining development was not regulated by the 1976 act but limited protection, regulation, and management had been achieved with the Shorelands Management and Protection Act (Michigan Legislature, Act No. 245, Public Acts of 1970) and the Soil Erosion and Sedimentation Control Act (Michigan Legislature, Act No. 347, Public Acts of 1972). Neither of these acts, however, specifically addressed non-mining development of the dunes. The Planning and Zoning Center study evaluated examples of state-run dune management programs in Georgia, North Carolina, and Washington as well as combination state and local management options. Recommendations included a variable 31 system in which dune development management was jointly exercised by local and state authorities. Following the passage of the Sand Dune Protection and Management Act (Michigan Legislature, Act No. 222, Public Acts of 1976) designated sand dune areas were delineated and companies were required to obtain permits in order to mine within them. Such permits must be renewed every 5 years. Mining activities within designated dune areas are regulated by the State of Michigan. Companies mining in designated dune areas must submit an annual report showing the area that is being mined as well as the amount of sand being removed from this area. Therefore, the accuracy of sand mining statistics (from coastal dune areas) has increased greatly since 1978. Mining of Michigan dune sand (since 1978) peaked in 1979, when over 3.3 million tons (3.0 million metric tons) were mined (Michigan Department of Environmental Quality, 2006; Figure 2.5). Between 1979 and 1989, the amount of dune sand mined fluctuated greatly, with a high of almost 2.7 million tons (2.45 million metric tons) mined in 1985 and a low of 1.5 million tons (1.36 million metric tons) mined in 1982 (Michigan Department of Environmental Quality, 2006; Figure 2.5). In six of the 10 years mining exceed 2 million tons (1.8 million metric tons) (Michigan Department of Environmental Quality, 2006; Figure 2.5). In 1989, the Sand Dune Protection and Management Act (Michigan Legislature, Act No. 222, Public Acts of 1976) was amended with Public Acts 146 and 147. The major change associated with these amendments was the creation of a new category of dunes known as Critical Dune Areas. Critical Dune Areas fall within designated sand dune areas and are less than 2 miles (3.2 km) from the lakeshore. They include the most 32 fragile and unique areas of the dunes. Critical Dune Areas have been mapped in the Atlas of Critical Dune Areas (Michigan Department of Natural Resources, 1989). Activities such as vegetation removal, earth removal, and construction require additional pennitting inside Critical Dune Areas. These amendments to the original act also decreed that no sand mining may be permitted within Critical Dune Areas. Exceptions to this rule are 1) the area was being mined prior to passage of these amendments, or 2) the proposed mining area is on property contiguous to a currently permitted mining operation for which sand mining rights were owned prior to the passage of the amendments. Approximately 80,000 acres (32,3 75 hectares) of Michigan’s 275,000 acres (111,289 hectares) of designated sand dune areas are Critical Dune Areas (29%). Coastal Dune Sand MIned In Michigan, 1978-2005 3,500,000 3,000,000 2,500,000 " 1,500,000 1,000,000 500000 0 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Year Figure 2.5: Annual tonnage of coastal dune sand mined in Michigan, 1978 to 2005 (data from Michigan Department of Environmental Quality, 2006). 33 The amount of dune sand mined generally increased between 1989 and 1994, with the peak for that period being over 2.6 million tons (2.36 million metric tons) in 1994 (Michigan Department of Environmental Quality, 2006; Figure 2.5). In 1994, the State of Michigan reorganized its environmental acts under the Natural Resources and Environmental Protection Act (Michigan Legislature, Act No. 451, Public Acts of 1994). This reorganization split the task of regulating and protecting Michigan sand dunes into two parts. Part 353 of the new act is entitled Sand Dune Protection and Management and manages all protection and management aspects relating to the dunes except sand dune mining. Part 353 is administered by the Michigan Department of Environmental Quality’s (MDEQ) Land and Water Management Division. Part 637 is entitled Sand Dune Mining and is administered by the MDEQ’s Geological Survey Division. Since 1994 the amount of dune sand mined has generally decreased, from approximately 2.6 million tons (2.36 million metric tons) mined in 1994 to 1.8 million tons (1.63 million metric tons) mined in 2005, the most recent year for which data is available. Major exceptions to this trend were the years 1998 through 2000, in which two of the three highest years on record for tonnage mined occurred: approximately 2.8 million tons (2.54 million metric tons) in both 1999 and 2000 (exceeded only in 1979) (Michigan Department of Environmental Quality, 2006; Figure 2.5). The average annual tonnage mined since 1978 is just over 2.2 million tons (2.0 million metric tons). Because dune sand is largely used by the auto industry, fluctuations in mining output may be related to changes in the economic health of that industry. Industrial sand production in areas outside of designated dune areas is unknown since these locations are not regulated by the state and thus are not required to submit annual reports detailing their activities. 34 Environmental Ethics, Sand Mining, and the Lake Michigan Coastal Dunes A large part of the impetus for reducing or eliminating the mining of coastal dune sand in Michigan is rooted in the history of the environmental ethics movement in the United States. As the movement emerged in the 19705 the rationale for saving coastal sand dunes in Michigan gained a new footing, finding support from a number of developing ethical ideas and theories. In order to understand the controversy surrounding sand mining in Michigan it is necessary to look at the issue in light of this movement. The emergence of environmental ethics as a distinct philosophical discipline is a relatively recent occurrence. One of the first works to call for an ethic of the environment was Routley’s (1973) piece entitled Is There a Need for a New, an Environmental, Ethic?. In this work, Routley (1973) argued that Western ethical traditions have focused on the implications of actions toward the agent acting and toward other people, but have largely ignored cases in which human interest is not at stake. Therefore, he proposed a new ethic that deals with humans’ relationship with the environment. In this same year, Naess (1973) introduced precursor ideas to the later- developed deep ecology movement. One such idea was ecological egalitarianism, which included the assignment of rights to living organisms and to non-living parts of the natural world. These ideas were later built upon by future environmental ethicists (e. g. Devall and Sessions, 1985). A major value theory formally described and supported by many environmental ethicists was that of intrinsic value in the natural environment. Rolston (1975) argued for an obligation to protect the integrity of ecosystems. He became a strong proponent of the belief that the natural world carries intrinsic value. Intrinsic value is a value theory in 35 which something has value just because of its own nature, irrespective of outside utility, beauty, etc. (Attfield, 2003). According to Rolston (1975), intrinsic value should be ascribed to species, ecosystems, and natural processes, and not just to individual organisms. He proposed the idea that humans have duties toward the natural world, such as protecting species and ecosystems from being destroyed. The emergence of belief in the intrinsic value of dune ecosystems (and ecosystems in general) has become one pillar of the argument against coastal dune mining. The theory of derivative value is in direct opposition to the notion of intrinsic value (Attfield, 2003). If something has derivative value then it derives value apart from its own nature. A commonly discussed type of derivative value is instrumental value. Instrumental value is value based solely on an object’s actual or potential usefulness (usually to human beings) (Attfield, 2003). Natural resources (including sand) by definition must contain instrumental value, since a resource is defined as something that is of use to human beings (Palmer, 1997). When a natural resource is used by human beings, it is instrumentally valuable to them. The ideas of intrinsic value of human beings and instrumental value to them are not new, but the growth of the environmental ethics movement has caused value theories to be applied in new ways. For example, value theories like intrinsic value were explicitly argued to apply to non-human entities (e.g. Rolston, 1975). By the mid-19705 the growing environmental preservation movement, based both implicitly and explicitly in the environmental ethics movement, created a new impetus for protecting the coastal dunes. The preservation of biodiversity, responsibility to future generations, rights of non-human organisms, and moral standing of the environment (e. g. 36 Russow, 1981; Goodpaster, 1978; Green, 1977; Singer, 1975) were all nascent ethical issues that had relevance to dune mining occurring in Michigan. In this context it was no coincidence that the Sand Dune Protection and Management Act (Michigan Legislature, Act No. 222, Public Acts of 1976) was passed by the Michigan legislature at this time. Today, appeals for coastal dune protection continue to list numerous aspects of dune systems that are perceived as “valuable”: their biodiversity, individual species, uniqueness, the ecosystems they contain, their use as a location for recreation, their beauty, and their provision of natural resources, to name just a few (Michigan Environmental Council, 2006). The field of environmental ethics has tried to explain why these items are construed as valuable. Taylor (1981) introduced a theory of value based on an ethic of respect for nature. This respect for nature is founded in a life-centered (biocentric) system of environmental ethics as opposed to a human-centered (anthropocentric) system of environmental ethics. According to Taylor, humans have moral obligations to Earth’s biotic community: to respect the integrity of ecosystems, to preserve species, to avoid pollution. All communities of life have a good of their own that can be affected for better or worse by humans. In order to respect nature it must be regarded as having inherent worth. Natural resource depletion, like most environmental ethical issues, has spurred ethical arguments that may be unresolveable because they are based on opposing theories of value (Harman and Arbogast, 2004). Harman and Arbogast (2004) considered the ethical values associated with the coastal dunes of Michigan and put forward a new rationale for preserving such ecosystems that avoids some of the bottlenecks that result from opposing value theories. They proposed that in situations where there are 37 conflicting claims of what is valuable the item(s) in question should be treated as valuable out of respect for those who hold them to be valuable. Thus, there are two parts of the proposal: to respect others, and to respect what they value. Harman and Arbogast (2004) proposed that respecting others finds its basis in our common humanness and that respecting what others value is not just part of respecting others, but is equivalent to respecting others. The policy and management implications of their new value system could be far-reaching. The central issue of any value dispute would become an empirical question of the degree to which the proposed uses depreciate the value of the resource to the stakeholders. Harman and Arbogast (2004) argued that when value disagreements occur preservation of a resource (in this case coastal dunes) should have preference over consumptive uses of a resource (e. g. mining dune sand) because preservation is the option that leaves open the most possibilities for future use. On the contrary, consumptive uses may permanently diminish or eliminate the source of value to some stakeholders. In summary, the emergence of the discipline of environmental ethics, in conjunction with the growth of the environmental preservation movement, has significantly affected public perception of Michigan’s coastal dunes and their consumption as a natural resource. New theories of value and of moral considerability have been put forth in the literature relating to central issues of the sand mining debate. However, most value judgments relating to the sand mining debate are not published explicitly in the literature (with their ethical basis discussed), and must be discerned based on public opinion and in publications that are not overtly dealing with ethical theories. Harman and Arbogast (2004) provided the sole study focusing explicitly on the 38 ethics of Michigan dune preservation and have presented a new rationale for preservation based on respect for others and what they value. Using their ideas it may be possible to make an empirical comparison of the value depreciation to stakeholders resultant from coastal dune mining versus value depreciation from inland options for mining foundry sand. Although such a comparison is beyond the scope of this study, this study can contribute to an eventual resolution of the dispute. 39 Chapter 3 STUDY AREA Michigan’s sand resources are largely a product of the state’s glacial history (Lewis, 1975) and are predominantly contained in sand dunes, outwash plains, beaches and beach ridges, lake-bottom sediments, and sandstone (Figures 3.1, 3.2). 92°W 90°W 88°W 86°W 84°W 82°W 0 50 100 Kilometers l—+——-l - Glacial outwash plains - Lake plains (sand and . - Modern alluviaideposits 44'" L — Beach Ridges 44'" - Water 42°N - 42°” 90°W 88°W 86°W M'W 82'W Figure 3.1: Potential non-bedrock Michigan sand sources (modified from Farrand and Bell, 1982 and Dorr and Eschman, 1970). 40 90°W 88°W 86°W 84°W 82°W I ' 0 50 100 Kilometers I—+——i 46°N _ 1: 46°N - tit» 44°N - 44°N 42°N » ‘ . 42°N co'w 88°W ee°w a4°w Figure 3.2: Michigan bedrock formations composed entirely or partly of sandstone. Most Lower Peninsula formations are buried by glacial drift and not exposed at the surface (data from Michigan Department of Environmental Quality, 1987). This study focused on interior Michigan; areas within 3.2 km (2 miles) of a Great Lakes shoreline were not considered. Detailed investigations and sampling were conducted on public land in 16 counties (Figure 3.3). 41 Figure 3.3: Counties in which sand samples were collected for this study Geology and Paleoclimate The current configuration of landforms in Michigan is due largely to its Pleistocene glacial history (Derouin et al., 2007; Schaetzl and Weisenbom, 2004; Larson and Schaetzl, 2001; Blewett and Winters, 1995; Rieck and Winters, 1993; Farrand and Eschman, 1974). In terms of bedrock geology, the Lower Peninsula of Michigan is dominated by the Michigan Basin, consisting of concentric beds of shales, limestones, and sandtones, the youngest of which are Late Jurassic in age (Dorr and Eschman, 1970). The bedrock of the Upper Peninsula is comprised largely of sandstones, shales, and 42 crystalline igneous and metamorphic units (Dorr and Eschman, 1970). Across the state of Michigan, except in the Upper Peninsula, bedrock outcrops are rare. In most parts of the state, bedrock i5 mantled by glacial drift that is tens of meters thick. In the northern Lower Peninsula this drift can be hundreds of meters in thickness (Rieck and Winters, 1993). This glacial drift is the result of at least six glaciations over the past 780,000 years (Larson and Schaetzl, 2001). In fact, due to reworking by subsequent glaciations, most drift deposits in Michigan only reflect the state’s most recent glacial episode, the Wisconsin, which lasted from approximately 110,000 years B.P. to 10,000 years B.P. (Larson and Schaetzl, 2001; Rieck and Winters, 1993). Across the 16 counties that are the focus of this study, landform development is largely related to three significant events: the retreat of the Wisconsin ice sheet (Rieck and Winters, 1993), climatic warming during the Hypsitherrnal period (Arbogast et al., 2002a), and variation in Great Lakes lake levels (Larson and Schaetzl, 2001; Thompson and Baedke, 1995). The first of these three significant events is the cause of most of Michigan’s large-scale landforms, such as outwash plains and moraines, while the latter two are responsible in part for the formation of coastal dunes, beach ridges, and lake plains, and the mobilization of inland dunes (Arbogast et al., 2002a; Larson and Schaetzl, 2001; Loope and Arbogast, 2000; Thompson and Baedke, 1995). Between 22 and 18 ka the Wisconsin ice sheet reached its maximum extent in the Great Lakes region. At about 18 ka, the Laurentide ice sheet began a general retreat northward across Michigan. This overall retreat was punctuated by several readvances, specifically at about 15.5 ka, 13 ka, 11.8 ka, and 10 ka (the Port Bruce, Port Huron, Greatlakean, and Marquette Stadials, respectively; Figure 3.4; Derouin et al., 2007; 43 Larson and Schaetzl, 2001). Glacial meltwater, produced in huge volumes as the ice sheet retreated, carried large quantities of sand and gravel that were deposited in fiont of the retreating ice in thick, stratified sheets of sand and gravel known as outwash plains (F arrand, 1988). 90°W 88°W 86°W 84°W 82°W 1 0 50 100 Kilometers i—I—i N . 46°N --' 15.5 ka eeeeeee118 ka 44°N ' ----I 10 ka 42°N - 90°W ‘ 88°W ‘ 86°W A 84°W 82°W Figure 3.4: Ice readvances in Michigan after the Wisconsin glacial maximum (~ 18 ka). During the readvances of 15.5 ka and 13 ka all of the Upper Peninsula was covered by glacial ice (modified from Larson and Schaetzl, 2001). 44 In the Great Lakes region, the Laurentide ice sheet developed a lobate pattern as ice channeled into valleys that would later become the Great Lakes. Three major lobes of ice affected Lower Michigan: the Michigan Lobe, the Saginaw Lobe, and the Huron-Erie Lobe. It was at the junctions of these lobes of ice that sand- and gravel-laden meltwater was most concentrated (Rieck and Winters, 1993; Farrand, 1988). After the Port Bruce Stadial (Figure 3.4) Allegan County was near the junction of the Michigan and Saginaw lobes of ice. After the Port Huron Stadial (Figure 3.4) the northern Lower Peninsula counties of this study, Antrim, Otsego, Kalkaska, Crawford, and Grand Traverse, were located in the Michigan-Saginaw interlobate area, sometimes called the north-central interlobate area (Rieck and Winters, 1993). Subsequently, these counties were covered with great thicknesses of glacial outwash (Schaetzl and Weisenbom, 2004). Wexford, Lake, and Newaygo counties have thick outwash deposits derived from meltwater of the Michigan lobe of the Wisconsin ice sheet. Ogemaw County contains outwash deposits from the Saginaw lobe of ice, while Lapeer and Tuscola counties were in the smaller Saginaw-Huron-Erie interlobate area. Alger, Schoolcrafl, and Chippewa counties, in Michigan’s Upper Peninsula, are located along the southern limit of the Marquette readvance (Derouin et al., 2007; Larson and Schaetzl, 2001; Blewett, 1994; Blewett and Reick, 1987; Figure 3.4). Stagnating ice after this advance is the likely cause of localized thick glacial outwash deposits found in these counties today (Derouin et al., 2007). Paleoclimate at the beginning of the Holocene was cooler than the climate found in Michigan today, but by approximately 9 ka the climate warmed, in what is known as the Hypsithermal period (Deevey and Flint, 1957). This relatively warmer and drier period, in addition to a drop in lake levels, allowed some inland sands to mobilize and 45 form dunes many tens of miles inland fiom today’s current lakeshores (Arbogast et al. 2002a; Arbogast et al., 1997). Subsequent cooling initiated stabilization and vegetation of these dunes. None of the state’s inland dunes are considered active today. Climate Michigan’s climate can be broadly characterized as humid continental. Much of the state is moderated by the surrounding Great Lakes (Michigan Legislature, 2001). Inland areas, away fiom this influence, have slightly more variability than their lakeside counterparts at similar latitudes. Average annual precipitation in the 16 counties that are the focus of this study ranges fi'om approximately 70 cm per year in Grand Traverse County, to over 100 cm per year in Allegan County. January average temperature ranges from -5.3° C in Allegan County to -9.6° C in Chippewa County. July average temperature ranges fi'om 21.7° C in Tuscola County to 182° C in Alger County (National Climatic Data Center, 2003). Snowfall is greatest in the Upper Peninsula, with some parts of Alger County averaging almost 400 cm of snow per year. Snowfall is lowest in Tuscola County, with less than 100 cm of snow annually, on average. Average growing season ranges from approximately 100 days in Alger County to almost 150 days in the southern half of the Lower Peninsula (National Climatic Data Center, 2003). Vegetation Vegetation across Michigan tends to fall into two broad categories: deciduous forests of beech, maple, oak, and hickory in the southern half of the Lower Peninsula and mixed forests (deciduous trees and conifers) of pine, spruce, fir, beech, maple, oak, and 46 aspen in the northern Lower Peninsula and the Upper Peninsula. The boundary between these two regions is known as the floristic tension zone (Schaetzl and Isard, 1991; Barnes and Wagner, 1981). Allegan, Lapeer, and Tuscola Counties are part of the deciduous forest zone of southern Michigan, yet they also contain areas of dry mixed forest, while other counties of the study area are majority mixed forests that are representative of northern Michigan (Schaetzl and Isard, 1991; Barnes and Wagner, 1981). Upland sites with sandy soils, common in the study area, are dominated by species that can tolerate dry conditions. In the southern part of the state such sites are dominated by oaks, while dry upland areas in the northern part of the state are dominated by pine, oak, and hemlock (Schaetzl and Isard, 1991). Soils Podzolization is the dominant soil process in sandy glacial outwash and inland dune areas of Michigan that are north of the floristic tension zone (Arbogast and Jameson, 1998; Arbogasteta1., 1997; Schaetzl and Isard, 1991). This process is accentuated by the adequate precipitation (especially snowfall) that occurs in much of the study area (Schaetzl and Isard, 1996). Dominant soils in the sandy outwash areas of the northern Lower Peninsula and Upper Peninsula are the Kalkaska, Rubicon, and Grayling series (Natural Resources Conservation Service, 2006). Inland dunes of the Saginaw Bay region commonly show Oakville soils. The Oakville series also often develops on sandy glacial outwash and dunes in southwest lower Michigan (Allegan county and surrounding areas) (Natural Resources Conservation Service, 2006). The dominant soil-forming process in sandy parent materials in southern Lower Michigan (generally south of the 47 floristic tension zone) is lessivage. These soils commonly have weak B horizons that are slightly enriched in clay (Schaetzl and Isard, 1991). 48 Chapter 4 METHODS In order to best locate alternative sand sources for the foundry industry I used a combination of mapping methods, field collection techniques, and laboratory procedures. The integration of geospatial analysis and traditional field and lab techniques allowed for a distinctly geographical approach to this problem. Summarized below are the methods used for this project. PRELIMINARY MAPPING METHODS There were many variables in the sample collection process for this project. These variables included determining a sampling scheme that would allow the most efficient and effective coverage of a very large study area, locating suitable sample collection locations according to this sampling scheme, accessing these sample locations, choosing an effective method for collecting the actual sand samples, deciding on an appropriate amount of sample to collect, and transporting the samples to the lab for further analysis. Each variable was addressed in a deliberate and calculated manner to ensure the best possible sample collection process given the time and resources available for the project. The first task of the sample collection process was to develop an effective and efficient sampling framework in order to best locate sites for sample collection. Other studies (e.g. Sundeen, 1978a) have essentially used a grid-like approach in which samples are collected across a very large region of interest, or a single-variable approach, in which 49 one variable is fixed (e.g. proximity to a railroad) and samples are collected and tested consecutively until another set of variables meets certain criteria. The approach used in this study minimized field sample collection by first narrowing the region of interest to only those areas most suitable based on economic criteria. To facilitate this effort a geographic information systems (GIS) package, ESRI’s ArcGIS 9, was utilized to spatially overlay data from surficial geology maps of Michigan (F arrand and Bell, 1982), USGS 7.5 minute topographic quadrangles, railroad network transportation maps, GAP Land Stewardship Maps (delineating Michigan public lands), Michigan Department of Environmental Quality (MDEQ) drinking water-well spot locations, Great Lakes shorelines, and critical dune areas (Michigan Center for Geographic Information, 2005). Associated with the water-well spot locations were lithostratigraphic data logged for each well. These data included the primary lithology of each stratigraphic layer encountered from the surface to the bottom of the water-well as the well was drilled, and the depth and thickness of each of these layers. While most data layers remained unaltered in the GIS, common database manipulation techniques were used to isolate and sum the thickness of those lithologic layers in the water-well logs that were, a) classified as sand or sand and gravel and b) within approximately 2 meters (7 feet) of the surface. The basic geospatial interpolation method of inverse distance weighting was then used with k equal to two and using the six nearest neighbors, to create a preliminary estimate map of areas of deep sand deposits that could potentially be sampled from the surface. The additional data layers identified above (e.g. railroad corridors, public land, etc.) were added to this preliminary sand depth map in the GIS. By spatially intersecting these relevant data layers, suitability maps were generated to aid in the selection of field 50 sample collection sites. Potentially suitable deposits (warranting a field visit) met the following criteria, 1) estimated thickness greater than 50 feet, 2) location within five miles of a railroad line, 3) outside of critical dune areas, and 4) on public land, to simplify sampling access. The locations of sand and gravel pits (public and private, as marked on USGS 7.5 minute topographic maps) were also digitized and added to the suitability map, as potential sample locations. FIELD METHODS Using the initial suitability map, potential areas for collecting sand samples were identified for field checking. Map-selected locations were evaluated in the field as to their suitability as sample locations. If surficial geology, public status, proximity to a railroad, location outside a critical dune area, and accessibility were confirmed, and no additional prohibitive factors were identified, the site was selected as a sample location. In all, 53 locations were selected for sample collection, in both the Upper and Lower Peninsulas (Figure 4.1). At each of the sample locations field data such as GPS coordinates, depth to water table (if applicable), sample depths, and other notable site and sample characteristics were recorded. Prior to the collection of samples from the aforementioned 53 locations, comparison field samples were collected at Wexford Sand Company’s Harrietta, Michigan, mine, which was found to be providing sand to Ford Motor Co. for foundry use. This sand urine is the only non-dune mine in the state currently providing foundry sand and has been operating since the mid-19705. For this reason, it was decided that sand collected at this location could serve as a proxy indicator of viable inland sand. 51 Samples collected at this location were used as a reference by which to compare other inland sand samples both physically and chemically. Reference samples were collected at four different locations within the active mining area, including three in-situ samples in the pit wall and one sample fi'om a talus apron. A fifth sample of processed sand was collected but is not included in the analysis. 0 50 100 Kilometers l—--l——-—l N e 0 Sample Location E] 8 km rail corridor 2 County H541 . 55133 Figure 4.]: Locations from which inland sand was collected and analyzed. The Wexford Sand Co. mine is labeled. A combined total of 167 samples were collected from the 53 locations sampled in this study. Most field sampling was conducted using a bucket auger. Samples were collected at 1.5 meter intervals to a depth of 6 meters. Thus, at a typical sample site, 52 samples were collected from depths of 1.5 meters, 3.0 meters, 4.5 meters, and 6.0 meters. If, for reasons such as stones, clay, or shallow water table, a depth of 6.0 meters could not be reached, samples were taken as permitted at the aforementioned intervals with a final sample being collected at the maximum depth reached. Sample volume collected at each interval was approximately one kilogram. This amount provided sufficient volume to conduct all necessary lab tests while still retaining sufficient reserves for any necessary retesting. Samples collected via bucket auger were collected directly from the bucket and immediately sealed in airtight plastic bags for transport to the lab. Several of the locations selected for sampling were pre-existing sand and gravel pits. At these locations sand was collected by hand at intervals determined by conditions at the sampling site. The locations were also visually inspected and described in order to facilitate comparison with the Wexford Sand Co. site and to better estimate the nature of deposits at which no exposures were present. Maximum depth for hand-collected samples was 15.3 meters below the surface. Minimum depth was 1.0 meter below the surface. Hand-collected samples were sealed and transported in a similar manner to bucket auger samples. At some locations a combination of hand-sampling and bucket augering was used. By typically collecting samples from four different depths at a single location, variability within a deposit could be better described and results could be rendered more representative of the deposit as a whole. Collecting samples up to 6 meters (or more) below the surface increased the likelihood of those samples being unaltered by near- surface soil processes. Such processes, while not affecting certain characteristics like 53 grain shape and sand grain size distribution, can alter characteristics like pH, acid demand, and clay content. Prior visual inspection of open exposures of inland sand deposits, including those at Wexford Sand Company’s inland mine, suggested that the chosen sampling scheme would allow for sufficient characterization of inland sand deposits given available time and resources. LABORATORY METHODS In order for sand—casted automobile parts to meet certain quality control specifications, sand used by the foundry industry must meet rigorous specifications in regards to numerous physical and chemical characteristics. Important characteristics for foundry sand are grain shape, grain size distribution, pH, acid demand, and 25-micron clay content. Requirements for these physical and chemical characteristics of foundry sand are described below. Grain Shape: For foundry casting, ideal grain shape represents a compromise between permeability, bonding ability, and smoothness of the finished surface of the casted part. Ideal foundry sand is generally subangular in shape. A subangular shape allows individual grains the ability to interlock sufficiently well to form a good mold or core while still providing necessary pore spaces for superheated gases to escape without breaking the mold during the casting process. A subangular shape also allows for a relatively smooth finished surface to the casted part. 54 Grain Size: An ideal grain size distribution of foundry sand is approximately normal and centered on US. Standard Sieve 70 (212 microns), with very little sand being retained on sieve sizes lower than 30 (590 microns) or greater than 140 (105 microns; Figure 4.2). In addition to these target values there is also a range of acceptable values for each sieve size (Table 4.1). Foundry Sand Grain Size Distribution % Retalned 0% 840 590420 300 212 149 105 74 : 53 <53 20 30 4o 50 70,100 140 200 270‘Pan Micron: (above) -- U.S. Standard Sieve Slze (below) Figure 4.2: General grain size distribution for foundry sand (from Ford Motor Co., unpublished). Specific users’ requirements can vary based on the casting technique used and the specifications of the finished product. 55 Table 4.1: General sieve ranges for foundry sand (from Ford Motor Co., unpublished). Specific users’ requirements can vary based on the casting technique used and the specifications of the finished product. Sieve # Microns Minimum % Maximum % Target % 20 840 0.0 0.0 0.0 30 590 0.0 2.0 1.0 40 420 4.5 13.0 10.0 50 300 25.0 35.0 32.0 70 212 30.0 40.0 36.0 100 149 10.0 20.0 16.0 140 105 2.0 6.0 4.5 200 74 - - 0.4 270 53 - - 0.1 Pan < 53 - - 0.0 Through 70 < 212 15.0 - - Through 200 < 74 0.0 1.0 0.1 pH: The target range for pH of foundry sand is 6.5 to 7.8. Values above or below this range can affect the function of chemical binders used in the casting process. Acid Demand Value: Acid demand value measures the chemical reactivity of the sand with an acid. Low acid demand values indicate minimal reactivity when mixed with an acid, and are desired for foundry sand. A threshold value of 10.00 was used in this research. Values above 10.00 can alter reaction speeds and affect the function of chemical binders. 56 2.5-micron Clay: Most foundry processes no longer use naturally occurring clay as a binder for molds and cores. For this reason, foundry sands will ideally have as little clay as possible. Tests of these physical and chemical characteristics were conducted according to the methodology that follows. Once the samples were collected they were subsequently air-dried to remove all moisture for at least 48 hours. Each dried sand sample (approximately 1 kg) was then homogenized using a sample splitter. In order to better characterize the target range of grain sizes, gravel, cobbles, stones, and large pieces of organic material (greater than 2mm) were removed from the sample. Laboratory procedures were performed on extracted subsets of these dried, homogenized sample materials. American Foundry Society (AF S) procedure AF S 1105-00-8: Sieve Analysis (Particle Size Determination of Sand) (American Foundry Society, 2004) was used to determine the grain size distribution of each sand sample. Approximately 50 grams of loose, dry sand was obtained from each sample and weighed to determine its exact weight to 0.01 grams. Once recorded, the sample was placed on a stack of the following sieve sizes (US. Standard): 6, 12, 20, 30, 40, 50, 70, 100, 140, 200, 270, and pan. The sieve stack was capped and agitated in an electronic sieve shaker for 15 minutes. After shaking the amount retained on each sieve was weighed and recorded. If the total weight 57 of recovered sand was not within 0.5% of the original weight the procedure was repeated with new sample. The results of procedure AF S 1105-00-8 were then used to calculate a grain fineness number for each sample, using procedure AF S 1106-00-S: Grain Fineness Number, AF S GFN, Calculation (American Foundry Society, 2004). The percent of each sample retained on a given sieve was multiplied by a sieve-specific multiplier (Table 4.2). These products were summed for each sample to determine an AF S Grain Fineness Number. Dominant grain shapes were classified according to procedure AF S 1107-00-S: Grain Shape Classification (American Foundry Society, 2004). AF S standard comparison grain shapes are split into four categories: rounded, subangular, angular, and compound (Figure 4.3). Approximately 15 sand grains were centered in the microscope field of view and using the AF S categories a dominant grain shape was noted. Grain shapes were also classified using the conventional geologic and engineering approach as outlined in Schoeneberger et al. (2002). Samples were pH-tested using an electronic pH meter immersed in a previously- agitated 2:1 water to sediment solution. Given the propensity for deep leaching of carbonates in sandy sediments, such as those collected, only the deepest of all samples from each location was pH-tested, in an effort to avoid anomalously low readings due to the pH-lowering effects of near-surface soil processes. pH—testing was also restricted to those sites in which samples greater than 3.0 meters below the surface were collected.- In addition to pH—testing, samples were also subjected to tests measuring acid demand. Acid demand value measures the chemical reactivity of the sand with an acid. 58 Table 4.2: Multipliers used to determine AFS Grain Fineness Number ASTM E—ll Sieve Size Multiplier 6 0.03 12 0.05 20 0.10 30 0.20 40 0.30 50 0.40 70 0.50 100 0.70 140 1.00 200 1.40 270 2.00 Pan 3.00 Compound Subangular Figure 4.3: AF S standard comparison grain shapes used in grain shape determination (modified from American Foundry Society, 2004) 59 High acid demand values can alter reaction speeds and the strength of binders used in forming molds and cores from sand. Acid demand tests were conducted on the same subset of samples that were pH—tested, since acid demand can also be affected by soil processes in a similar manner. Tests were done according to American Foundry Society procedure AF S 1114-00-8 (American Foundry Society, 2004). Because particles smaller than 25 microns in diameter can have a significant effect on the permeability of a mold or core, and thus on the finished surface of casted products, all particles below this threshold level are described by the AF S as 25-micron clay (American Foundry Society, 2004). The larger end of this range (3 to 25 microns) is considered silt according to the USDA sediment fractionation scheme (Schoeneberger et al., 2002). The 25-micron clay content of all collected samples was determined using a modified version of the AF S hydrometer method (American Foundry Society, 1978). Samples were dispersed for at least 10 hours in a medium-speed agitator and then brought to standard temperature (20° C). Using Stokes’ law, the settling rate was determined for particles 25 microns in diameter. Suspended sediment measurements were taken in a 1- liter graduated cylinder using an ASTM 152H hydrometer, at approximately 5 minutes. Results are in 0.5 percent increments. MAPPING SAND THICKNESS Michigan drinking water-well logs, each containing stratigraphic information from the surface to the depth of the well, were used to estimate sand thickness. Water- well data were obtained from the Michigan Center for Geographic Information. Data 60 were initially compiled by the Michigan Department of Environmental Quality based on required submissions by well-drilling companies (Michigan Center for Geographic Information, 2005). The geostatistical technique of ordinary kriging (Venteris, 2007) was used to interpolate sand thickness information between locations at which this information was known (i.e.water-well locations). Common database transformation techniques were first used to derive, for each well, two sand thickness statistics: 1) the cumulative thickness of all “sand” stratigraphic layers between the surface and the bottom of the well (“sand” layers, in this context, are those layers in which sand was listed first in the stratigraphic layer name and silt and/or clay were not present, i.e. “sand,” “sand & gravel,” “sand & cobbles,” “sand & stones,” or “sand & boulders”); and 2) the thickness of sand within 23 m (75 feet) of the surface. Once these calculations were made the data were filtered numerous times to assist in removing any bad data points. Some examples of indications of a data point being unsuitable were, 1) more than 75 feet of sand was listed (in the well log) within 75 feet of the surface, 2) the listed depth of a stratigraphic layer was greater than the listed depth of the well, 3) a layer of bedrock was listed with glacial drift material below it. The number of water wells used in this analysis was reduced from approximately 327,000 to about 318,000 afier removing questionable data. ‘ In order to minimize the effects of nonstationarity in the data, each county was kriged individually and then the results were mosaicked together. Using ESRI’s ArcGIS 9, county boundaries were buffered 10 km and then used to clip the Michigan water-well data set (Michigan Center for Geographic Information, 2005). For each county, kriging was performed on all the water-well point locations within the county or within ten 61 kilometers of its boundary. Spatial dependence was modeled on an individual county basis. For most counties, a spherical model was used for the semi-variogram with the nugget, sill, and range values being determined automatically by ArcGIS. Five neighbors were used in each of four sectors, divided on the diagonal around the location being interpolated. If the data appeared to indicate anisotropy, new parameters taking this into account were generated automatically based on the nature of the data. The kriging output was clipped to the actual county boundary. The overlap of wells used for each county assured that there would be a minimum of edge effects in the data once it was mosaicked together for the whole state. The resulting mosaic of sand thickness estimates for Michigan’s 83 counties provides coverage for almost the entire state. Most Great Lakes islands are not included in the dataset due to a lack of data points for these areas. In addition, interpolation (extrapolation) was not extended to areas lying outside the bounds of the data. For this reason, a few areas of the Upper Peninsula do not contain sand thickness estimates. 62 Chapter 5 RESULTS AND DISCUSSION The purpose of this study is to determine whether suitable inland sand sources exist in Michigan that could be utilized by the foundry industry as an alternative to coastal dune sand. The utility of inland sand is dependent upon many factors, including 1) size of the deposit 2) accessibility to markets, 3) grain shape, 4) grain size distribution, 5) 25-micron clay content, 6) pH, and 7) acid demand. This chapter describes the results of the economic, physical, and chemical analyses that were performed to determine the viability of Michigan’s inland sand resources for the foundry industry. ECONOMIC ANAL YSIS Physically and chemically superior sand is of little value if it cannot be quickly and cheaply brought to market. In this context, two important economic factors considered by the foundry industry are 1) the volume of a sand deposit, and 2) the accessibility of the site. Every potential site must contain a sufficient volume of sand to allow for operation for along period of time, in order to recoup start-up costs. Accessibility is critical because of its direct impact on the cost of transporting sand from mine to market. Although this is a simplified view of the economic factors that are involved in sand mining, these two components are the most important in determining the economic viability of inland sand resources for the foundry industry (Marrone, 2005, pers. com.). The following discussion outlines the issues associated with these two variables and their importance to this study. 63 Transportation Costs The cost of transporting sand from a deposit to its point of use can significantly affect the total cost of the final delivered sand product. Therefore, reducing transportation costs is one means of making inland sand deposits better suited to foundry and other industrial uses. The ideal method of moving large amounts of sand cheaply to market is barge transportation (Marrone, 2005, pers. comm.). However, for barge transportation to be used as the sole means of transport the deposit must be located in close proximity to a navigable waterway. Very few barge-navigable inland rivers exist in Michigan, leaving the Great Lakes as the only major option for using this method of transportation. Because of this lack of navigable rivers in Michigan the sole use of barge transportation is incompatible with the use of inland sand. The second-best method of transporting foundry sand to market is by rail (Marrone, 2005, pers. comm). Michigan contains an extensive network of rail lines which could be used to transport sand from interior locations (Figure 5.1). Ideally, potential inland sand deposits would be located within 8 km (5 miles) of an existing rail line (Marrone, 2005, pers. com.). This geography would allow for quick and easy construction of a spur line to access the deposit, reducing significantly the need for costly investment in new rail networks or expensive truck transport. Much industrial sand is currently transported in this manner. Approximately 64,674 kmz, almost exactly one-third of the state, lies within 8 km (5 miles) of an existing rail line (and more than 3.25 km (2 miles) from the lakeshore) (Figure 5.1). Most areas that lie within 8 km (5 miles) of an existing rail line are in the 64 southern half of the Lower Peninsula. Rail lines connect all major cities in this part of the state and are especially dense in the Detroit-Saginaw corridor. Rail lines in the northern Lower Peninsula connect larger cities to the more extensive network to the south. They extend to Ludington, Manistee, Cadillac, Traverse City, Petoskey, and Alpena and generally run in a north-south direction. Although the Upper Peninsula contains railroad lines, they do not cross the Straits of Mackinac. This geography means that sands cannot be transported by rail directly from the Upper to the Lower Peninsula but must be moved by barge at some point. Ice buildup and the closure of the Soo Locks prohibit barge transport on Lake Superior during the winter months of the year (US. Army Corps of Engineers, 2007). Such a route, however, could be used during the summer. x “J. (W ‘ . / .1 t \\‘l m, \ ultSte.Mane E I «b‘ x\ l \xv g n», l at Capital L a Major City \ - Rail Corridor “. he \ k - \ 4/ County K \ Cl “C" . < (A, 0 40 80Kilometers r”) ( . gi }_+_.{ N :1? )—lr\t > \ ‘ ' l e “ ' ' V Q l ’ .. \ .. 5:?“ A“ ,1 , Detroit / _, .7 3 A. f i V, 1 / Figure 5.1: All areas greater than 3.25 km (2 miles) from a Great Lakes shoreline within 8 km (5 miles) of an existing rail line. 65 Results of this transportation analysis suggest that sand in Michigan’s Lower Peninsula, especially its southern half, would be easiest to access, and thus likely the cheapest to bring to market. However, inland sand fiom Michigan’s Upper Peninsula cannot be ruled out as an accessible source of sand for the foundry industry because a combination of rail and barge transport could be used. A more in-depth study focusing solely on transportation costs would be necessary to obtain specific cost estimates for transporting sand from different regions of the state. Such an analysis could determine whether the physical location or the physical and chemical characteristics of a sand deposit have a greater affect on the cost of the delivered sand product. The results of the current analysis, however, do identify the best areas of the state for inland sand mining that avoid the need to make significant changes to the current transportation infrastructure. Sand Thickness and Volume The estimated volume of sand in a given area is a crucial economic factor in a sand deposit’s viability for use by the foundry industry. A geostatistical analysis, using data from water-well logs, has been used to identify those general areas of the state where great thicknesses of sand are most likely present. The overall accuracy of the results is dependent on many variables, including the accuracy of the water-well data used, the amount of spatial dependence exhibited by the data, and the scale at which the data are intended to be used. Accuracy was improved by filtering the water-well data to remove bad data points prior to running the geostatistical model, as described in Chapter 4. 66 Spatial dependence was modeled more accurately by kriging each county individually and then mosaicking the results together. Because slight changes in water well lithology (e.g. sand and gravel vs. sand, clay and gravel) can sometimes cause nearby wells to have significantly different sand thickness values, variance in the data was fairly high. Mean standard error by county for cumulative sand thickness was generally between 6 and 12 meters. Mean standard error by county for near-surface sand thickness was generally between 4 and 8 meters. Sand thickness results are presented for general planning purposes, to show the general trend in sand deposit thickness (and therefore volume) across Michigan. It is inappropriate to use them for predicting the exact sand depth (or thickness or volume) at any specific location. This analysis shows the broad trends across Michigan of a minimum estimate of sand thickness (Figure 5.2). These results are considered a minimum estimate for sand thickness because many of the wells used in the analysis are drift wells (tapping an aquifer contained in glacial drift) and therefore do not reach bedrock. For any drift well, it is possible that even more sand is located below the level to which the well was drilled. Water-wells clearly indicate that the thickest sand deposits in Michigan are in the northwest portion of the Lower Peninsula (Figure 5.2). This is in agreement with Rieck and Winters (1993), who also estimated that the greatest total drift thickness is in this part of the state. The cumulative thickness of sand between the surface and bedrock in most of this region is estimated to be at least 20 m, with some areas likely having cumulative sand thicknesses over 75 m. The thickest deposits are largely located in the north-central interlobate area (Rieck and Winters, 1993). This area was between the retreating Michigan and Saginaw lobes of ice during the late Pleistocene and thus became a 67 50 100 Kilometers Minimum Estimate Cumulative Sand Thickness (m) -0-5 -5—10 -1o-15 -15-20 -20+ l:lCounty Figure 5.2: Minimum estimated cumulative sand thickness between the surface and bedrock. Data are minimum estimates because not all water-wells used in the estimation process penetrate to bedrock. 68 repository for large amounts of sandy glaciofluvial material (Blewett and Winters, 1995; Rieck and Winters, 1993). Other significant concentrations of sand occur in the southwest Michigan interlobate area and scattered across the length of the Upper Peninsula. Areas estimated to contain very little sand include much of the southeast quarter of the Lower Peninsula, the extreme northeastern portion of the Lower Peninsula, and areas of the Upper Peninsula where bedrock is very shallow or even at the surface (Figure 5.2). A more realistic look at potentially accessible sand deposits is possible by mapping sand thickness within 23m (75 ft) of the surface (Figure 5.3). Because glacial drift in the northwest portion of the Lower Peninsula can be over 75 m thick, much of the sand detected during well drilling is likely not feasible to mine at present. Each potential location would need a detailed survey to determine exactly how deep sand could feasibly be extracted. Areas estimated to have thick deposits of sand within 23 meters (75 ft) of the surface are considered more economically accessible than areas with a cumulative sand thickness of 23 m (75 it) between the surface and bedrock. Technical advances in sand mining may allow deeper sand deposits to be accessed in the future. Areas with the most near-surface sand are largely located in the north-central portion of the Lower Peninsula (Figure 5.3). Additional areas with large amounts of near-surface sand include Muskegon, Lake, and northern Newaygo Counties in the Lower Peninsula and Alger, Schoolcraft, and Luce Counties in the Upper Peninsula. In certain areas, such as Emmett County, results suggest large amounts of total cumulative sand (Figure 5.2), but much smaller amounts of near-surface sand (Figure 5.3). In such instances, much of the sand in the deposit is likely located well below the surface (at least 69 50 100 Kilometers Estimated Cumulative Sand Thickness (m) Upper23m - 0-5 - 5-10 - 1o- 15 -15-20 - 20+ E County Figure 5.3: Estimated cumulative thickness of sand between the surface and a depth of 23 m (75 ft). 70 below 23 m) and therefore may be difficult to extract. In terms of economic viability, the volume of near-surface sand is a more important statistic than the total cumulative volume of sand in the entire profile. Because kriging is a method of interpolating between known data points to estimate unknown values, the best results are in areas where well density is high. In those areas with many closely spaced wells that show strong spatial dependence, estimates are likely more accurate than in those areas where there are few wells and/or where spatial dependence between them is low. The highest concentrations of water wells are in the southern half of Michigan’s Lower Peninsula (Figure 5.4), giving higher confidence to the sand thickness estimates for this part of the state. The lowest concentrations of water wells are in rural areas of the Upper Peninsula and, accordingly, confidence in sand thickness estimates in these areas is lower. Other contingencies beyond well density also affect the sand thickness estimates and are addressed below. Results of the sand thickness analysis provide minimum estimates for sand thickness, but by mapping which wells are bedrock wells and which are drift wells it is possible to predict those areas in which even more sand may be present than the stated minimum estimate (Figure 5.5). In areas of the state where most wells are drift wells there is a high likelihood that sand volume is higher than estimated. This includes much of the north-central Lower Peninsula and, to a lesser degree, the southwest comer of the Lower Peninsula. Areas in which most water is supplied by a municipal water system, such as the Detroit metropolitan area, also have few water wells (Figure 5.5). In areas in which bedrock wells are predominant, sand thickness estimates will be more accurate because it is not possible for sand to be present below the depth of the 71 50 100 Kilometers Water Wells by public survey township - 0 -1-10 - 11 -100 - 101 -500 - 501 -1.000 - 1,001+ - Nodata : Township U County Figure 5.4: Number of water wells, by public survey township. Public survey townships are usually 36 square miles in area, 6 miles on a side. 72 0 50 100 Kilometers I---l—I N e: . fi 0 o. . . N. -t , . 0 M n. 0'. 'l K ’ 0 '3. \ Figure 5.5: Michigan bedrock wells. Drift wells are not shown. Bedrock wells are common in the southern half of the Lower Peninsula and in the Upper Peninsula, but are virtually absent from the northern Lower Peninsula. 73 well. Bedrock wells are predominant in southeast and south-central Lower Michigan, the Saginaw Bay region, the “Thumb” region (tip of the Huron Peninsula), and much of the Upper Peninsula (Figure 5.5). In these areas sand thickness estimates will be closer to their actual values, while in drift well areas sand thickness estimates likely underestimate the actual sand thickness. Areas estimated to contain the greatest sand thicknesses (the north-central Lower Peninsula) are also the areas where sand thickness is most likely underestimated. Drift wells are often completed in sand because sand is an excellent aquifer. Most water-wells in the northern half of the Lower Peninsula are drilled into a water-bearing unit of sand (Figure 5.6). Many wells in southwest Michigan and a smaller number in southeast Michigan also have sand as their lowest stratigraphic unit (Figure 5.6). It is not unreasonable to assume that this lowest sand unit extends below the depth of the water well. When this is the case, sand thickness is underestimated. Most drift wells footed in sand are in areas estimated to have high sand thicknesses (Figures 5.2, 5.6). Again, those areas most likely to have the greatest amounts of sand are also those areas most likely to be underestimations of the actual sand thickness. Inland Michigan contains extremely large deposits of sand that tend to be located in glacial interlobate areas. The thickest of these deposits is in the north-central portion of the Lower Peninsula, with secondary large volumes likely in the southwestern portion of the state. Estimated thicknesses of sand deposits in north-central Lower Michigan exceed 75 m in some areas and can spatially extend over many square kilometers. In southwest Lower Michigan estimated thicknesses exceed 30 m in small, localized areas. In the Upper Peninsula, sand deposits are widely scattered with shallow bedrock being 74 extensive in most areas that lack deep sand. Localized deposits have been estimated to exceed 45 m in thickness, but the relative lack of wells on which these estimates are based leads to lower confidence in results for this area of the state. In those areas found to have thick sand deposits, drift wells, as opposed to bedrock wells, are much more prevalent, with the sand being a water-bearing unit. Because these water-bearing sand units continue below the depth of the well it is these areas of thick sand where the total sand thickness and volume have most likely been underestimated. 50 100 Kilometers Deepest Stratigraphic Unit in Well Log 0 Sand 0 Not Sand Figure 5.6: Lowermost stratigraphic unit of Michigan water-wells. Locations in pink are wells completed in sand, while those in gray are wells completed in bedrock or non-sand drift. 75 PHYSICAL ANAL YSIS 0F SANDS Important physical factors that determine whether sand can be used in the foundry process include grain shape, grain size distribution, ‘25-micron clay’ content, refractoriness, and moisture content (American Foundry Society, 2004). Grain shape, grain size distribution, and ‘25-micron clay’ content have been tested explicitly in this analysis. The latter two factors were not tested for the reasons listed below. Refractoriness refers to a mineral’s ability to withstand firsion due to high temperatures. For a mineral to be used in casting it must be able to withstand temperatures hotter than the melting point of the metal being cast without fusing. Iron melts at 1535° C, steel at 1510° C, and aluminum at 650° C. Silica, the primary component of coastal dune sand and inland sand in Michigan, has a fusion point of 1710° C. Therefore, it will not fuse when casting iron, steel, or aluminum. Because Michigan sand is composed primarily of silica (as opposed to zircon sands, for example, which also can be used for casting molds and cores) it is assumed to have satisfactory refractoriness for the common types of foundry casting in Michigan. Moisture content of the sand, while crucial in the foundry process, can easily be artificially manipulated prior to beginning the casting. For this reason, it was unnecessary to test the in situ moisture content of inland sand samples. Results of the grain shape, grain size distribution, and 25-micron clay analyses are described in the following sections and have been grouped by general geomorphic setting, either inland dunes or glacial outwash. Glacial outwash is widespread across Michigan but can adjoin or overlie texturally similar sandy lacustrine sediments or sandy till in some areas. A detailed differentiation of such deposits was not within the scope of 76 this project. Thick deposits of sand identified in this project may be a mix of outwash sand, sandy till, and/or sandy lacustrine sediments. Samples used in both the physical and chemical analyses were collected from the locations shown in Figures 5.7a, b, and c. Each of these figures shows sample locations in specific sub-regions of Michigan. Gladwin Bay Saginaw Montcalm Clinton lngham Calhoun Jackson Kalamazoo \ Hill§ale / Figure 5.7a: Sample locations in southern and central Lower Michigan. 77 Emmet Charlevoi 0 l--—1—l Cheboygan Presque Isle W 5690 Montmorenc Alpen Obego 2 1 rd 1 Oscoda Alcona ord 2 gemaw 2 I O osc e gem 09 aw J) gr Antrim G d Antrim ’ ' 1 Antr 2°/ Leelanau * r g. Benzie Grancf/ { ' raverse JGrandT ve e1. . 9 Wexf d Missaukee Manisteey o ow 0rd5 ord-t rd1 lwe ord ”T \Wexyr an o. ‘ “ \ 350" .Lake 03* \c... l are “I 7x Gladwin . ren Figure 5.7b: Sample locations in northern Lower Michigan. 78 I O 15 30 Kilometers l—+——-i N \ , _ < 1 7 ' 1 l X -‘ ’ ~ , ' Luce \ ~ 3 I —C_hipp-.. Ch W Chippewa 4 ‘ -- (d? -_ , M . Th e " % 8 I ‘ pg . /' i I Emmet 1 5) l ‘ ’ ’ , Chaigvoi Figure 5.7c: Sample locations in the Upper Peninsula. 79 Grain Shape Grain shape is important in the casting process because it can affect the permeability and strength of the mold or core. Ideal grains are subangular in shape to provide a compromise between the added strength of interlocking grains and the permeability of open pore spaces. Of Michigan dune sand users surveyed by Ayres and Chapman (1978), approximately 70% used subangular grain shapes or had no grain shape requirements. To determine their shape, 145 samples from 47 glacial outwash locations and 20 samples fiom 6 inland dune locations were analyzed with a microscope, generally at 25x and 40x. Using approximately 15 sand grains in the central field of view, samples were assigned a dominant grain shape. They were categorized using both the American Foundry Society (AF S) matrix (see Figure 4.3) and the conventional geologic and engineering grain shape matrix as presented in Schoeneberger et al. (2002). The latter will henceforth be referred to as the conventional matrix. Additional samples were also analyzed for comparative purposes from the Wexford Sand Co. mine (Figure 5.7b). Grain shapes varied little across all locations, with no difference between glacial outwash samples and inland dune samples. Using the AF S 4-group categorization matrix (rounded, subangular, angular, or compound) approximately 96% of all samples had a subangular grain shape. Only 2 samples had compound grains, while 3 samples had angular grains. Using the conventional 30-group matrix, approximately 75% of all samples had a subangular roundness and a sub-discoidal sphericity (Figure 5.8). Approximately 8% of samples were classified as angular using the scheme, while 11% were classified as 80 subrounded. Although the conventional scheme is more specific in its classifications, ultimate suitability determinations were made using AF S classifications. Roundness Very Angular Angular Subangular Subrounded Rounded Well Rounded Q O O O 0 Sub- discoidal 0.5 Discoldal 1.5 Spherical 2.5 Sphericity woeoa: Sub- prismoidal 3.5 enema DQGflQ= Prismoldal 4.5 wQQBS: ®©O©Q Figure 5.8: Conventional geologic and engineering grain shape classification matrix with sample results plotted. Ideal grain shape for foundry use is subangular (modified from Schoeneberger et al., 2002). Glacial Outwash Locations Approximately 96% of all outwash locations had grain shapes that were subangular using the AF S scheme (Appendix C; Figure 5.9). Eighty-three percent of outwash locations had subangular roundness using the conventional classification matrix (Figure 5.10). Examination revealed that the smallest grain sizes (<100 um) tend toward being angular in shape, while in many cases grain sizes larger than approximately 400 L:’fi Hi .ng Sample Locations A Inland dune 0 Glacial outwash AFS Grain Shape J \flj’p/ . CA All samples subangular / ., ,1 o A 1 or more sample not (w 1:9/ / ‘\ ) subangular 1 Ir , - Railroad Corridor 1 E :\ l ._ J‘ \I \\ in; '. [3 County Boundary 73“ ‘ K \T‘\ “I A“ * Wexford Sand 00. __ g. \f\ ‘r: ,\ I : l l l \\ . l \ ‘ j"z.’1 “ < fif‘l—L‘; . Z \ \ .rtlz ‘T5fl ‘» *7 r r» ,\ 1 , 1 ‘ h \ ,1 l l“ , ,. Amt ‘ ,‘i 1‘ ’ F A}. L o. , \\._ . _. 1, - . \ 1 ." l \j’ l / J) .. , ‘ / l K 4 \ , 4 A 7 —v‘—‘ {1 A“ \ / \j Figure 5.9: AFS grain shape results. The predominant grain shape was subangular in over 95% of the samples. 82 Roundness Very Angular Angullr Subangular Subrounded Well Rounded 0.5 1.5 2.5 35 5.5 ”QQGaOO discoidal 0.5 1.5 A? E s herical s {30000 O: m Sub- 3.5 4.5 Figure 5.10: Conventional geologic and engineering grain shape classification matrix for glacial outwash samples. Ideal grain shape for foundry use is subangular (modified from Schoeneberger et al., 2002). microns were more rounded (Figure 5.11). Since the majority of sand grains were between 100 and 400 microns in size the dominant grain shape was almost always subangular. Sand currently being extracted from the Wexford Sand Co. mine was also analyzed for grain shape. This sand is derived from glacial outwash and was predominantly subangular in shape, having a conventional roundness value of 2.5 and a sphericity value of either 1.5 or 3.5 depending on the sample (median rho values; see Folk, 1955). All samples from this location appeared very similar (in terms of grain shape) to most samples collected from other inland locations (Figure 5.12). 83 Figure 5.11: Outwash sand grains from Crawford County (Crawford 4, sample C). Predominant grain shape was largely subangular, but grains larger than approximately 400 microns in diameter often showed distinct rounding. Arrows identify more rounded grains. Magnification is 40x. 84 Figure 5.12: Sand from Wexford Sand Company’s inland sand mine, 25x magnification. Notice the increased roundness of the larger grains but the predominance of subangular shape. A small number of glacial outwash sites had samples that failed to meet AF S grain shape criteria for foundry usage. Two locations, Wexford 4 and Wexford 1, exhibited compound grains, which are unacceptable for use by foundries (Figure 5.13). At Chippewa 6 both samples had angular grain shapes. In three samples, sand grains were obscured by silt grains and clay particles and grain shape could not be accurately determined (see Appendix C). Obscured sand grains were also an indication that clay content was likely far too high to be used unprocessed for foundry castings. Less than 5% of all outwash samples had grain shapes that were not classified as subangular using the AFS matrix. 85 Figure 5.13: Sand sample from a depth of 6 m (sample D) at Wexford l. Arrows point to compound sand grains. Magnification is 25x. The gain shapes of glacial outwash samples showed strong vertical consistency at most locations. In no instance did samples from different depths at the same location vary by more than one unit of roundness (conventional matrix). Twenty-one glacial outwash locations had a single predominant gain roundness common to all samples. Most glacial outwash samples had sphericity values of 1.5 (median rho values; see Folk, 86 1955). Sand gains with sphericity values of 2.5 or 3.5 were usually still present, but in much smaller amounts (see Appendix C). Few outwash locations failed to show the general consistency of grain shape that was common overall. Wexford 1 (Figure 5.13) and Wexford 4, which exhibited compound gains at 6m (sample D) and 1.5m (sample A) respectively, did not have compound gains at any other depths. Inland Dune Locations Inland dune samples (Table 5.1) showed little difference from glacial outwash samples with regard to gain shape (Figure 5.14a, b). Approximately 96% of inland dune samples had subangular gain shapes measured using the AF S matrix. This is consistent with glacial outwash results. Using the conventional matrix, inland dune samples showed slightly less variability of grain shape than glacial outwash samples (Figure 5.15). Similar to outwash samples, inland dune samples also showed increased rounding with larger grain sizes. All inland dune samples, except for one, met the necessary gain shape criteria for foundry use. Chippewa 3 (Figure 5.7c) had angular grains at a depth of 4.5 m (sample C). All other locations had all subangular gain shapes. Inland dune locations, like glacial outwash locations, also showed strong vertical consistency. Conventional roundness never varied by more than one unit at any location. All samples had a predominant sphericity value of 1.5 (Table 5.1). 87 Figure 5.14a: Sand from an inland dune site in Arenac County (Arenac 2, sample D). Magnification is 25x. ' ‘7 :~.: as... " ‘ ‘V'a 3 _ g, m} g .- is "I ,5? .-a' :5. > k , _ a, ,- Figure 5.l4b: Glacial outwash sand from Crawford County (Crawford 2, sample D). Although average grain size is slightly larger than at the inland dune site in Arenac County (Arenac 2), grain shape is virtually indistinguishable. Magnification is 25x. 88 Table 5.1: Conventional and AF S grain shapes for inland dune samples. Location Sample Conventional Grain AFS Grain Shape Roundness 2.5 3.5 2.5 2.5 2.5 3.5 2.5 2.5 3.5 3.5 2.5 2.5 3.5 1.5 2.5 1.5 2.5 2.5 2.5 2.5 2.5 2.5 Arenac 2 Arenac 2 Arenac 2 Arenac 2 C om>ow>ow>oom>om>oom> 1 1 1 2 2 2 2 3 3 3 4 4 4 5 5 5 89 Roundness Very Angular Angular Subangular Subrounded Rounded Well Rounded 35 4.5 5.5 Sub- discoidal Discoidal 5% 1.5 .E‘ if 2 {3 D O O O O =- U) 3.5 Prismoidal 4.5 Figure 5.15: Conventional geologic and engineering grain shape classification matrix for inland dune samples (modified from Schoeneberger et al., 2002). Grain Size Distribution An ideal gain size distribution of foundry sand is approximately normal and centered on US. Standard Sieve size 70 (212 microns; Figure 4.2). Very little sand should be retained on sieve sizes lower than 30 (590 microns) or geater than 140 (105 microns; Table 4.1). In order to represent a gain size distribution with a single number the American Foundry Society calculates a grain fineness number (GFN) based on the percentages of sand retained on each sieve used in the sieving process. Grain fineness numbers (GFNs) of acceptable foundry sand fall between 47 and 53. Higher GFNs sigiify finer sands, whereas lower GFNs signify coarser sands. In addition to a target 90 range for GFNs there is also a range of acceptable values for many screen sizes (Table 4.1). To be ideal, sand used by the foundry industry must not only have a GFN that falls within the target range, but must also have an acceptable percentage of sand retained on each sieve or group of sieves (Table 4.1). Grain fineness numbers measured in this study ranged from a low of 32.2 at Wexford 6 to a high of 171.5 at Chippewa 6. The median GFN of all samples was 48.8 and the mean GFN was 53.8. A total of 38 samples had gain fineness numbers that fell within the target range of 47 to 53. These 38 samples were from collected from 24 different sample locations, an indication of the high amount of variability in GFN results (Figure 5.16). Most sample locations from the Upper Peninsula had no samples within the target range. Other areas in which few samples fell within the target range included Lapeer and Tuscola Counties, eastern Wexford County, and Otsego County. Glacial Outwash Locations Both the median GFN (48.2) and the mean GFN (52.5) of glacial outwash samples fell within the target range. However, the overall range of GFN for glacial outwash (32.2 to 171.5) was large (see Appendix C). Grain size distribution is a much more variable characteristic in glacial outwash sand than is grain shape. Only at Wexford 1 (Figure 5.7b) did all samples from a location fall within the target range. All other glacial outwash locations had at least one sample outside of the target range. The variability in grain size distribution is not surprising given the fluvial origin of these deposits. Fluvial deposits are typically well stratified, with fining upwards sequences and various degrees of sorting in the sediments. As a result of this variability, 37 glacial 91 0 30 60 Kilometers l—i—l Sample Locations A Inland dune 0 Glacial outwash Grain Fineness Number 0 A 1 or more samples meet target 0 ‘ No samples meet target - Railroad Corridor [:] County Boundary * Wexford Sand Co. Figure 5.16: Grain Fineness Number (GFN) results. Target range for foundry sand is 47 to 53. Labeled locations have one or more samples within the target range. 92 outwash samples that met the target GF N criteria were from 23 different locations (Figures 5.16 and 5.17). Many locations only had one or two samples that were within the target range. Samples taken fi‘om Wexford Sand Company’s inland sand mine showed similar GFN variability. The samples collected from this location had grain fineness numbers that ranged fi'om 41.4 to 49.4. The standard deviation of these four samples was 3.83. Using standard deviation it is possible to make comparisons between this mining location and other sample locations on the basis of grain size variability (Figure 5.18). Standard deviation of grain fineness numbers for glacial outwash sand ranged from 1.20 to 41.01 and is a good measure of grain size variability at a location (Figure 5.18). The median standard deviation was 5.85 and the mean standard deviation was 8.47. Therefore, on average, these locations express slightly more variability than would be expected at Wexford Sand Company’s mine. However, 12 glacial outwash sample locations exhibited less variability in grain size than the sampled deposits at the Wexford Sand Company. These locations were Allegan 4, Alger 4, Antrim 2, Chippewa 6, Crawford 1 and 4, Grand Traverse 2, Lake 1, Newaygo 1, Otsego 3, and Wexford 1 and 6 (Figures 5.7a, b, c). In addition, three glacial outwash locations, Grand Traverse 2, Newaygo 2, and Wexford 1, had both less variability in grain size than Wexford Sand Company and at least one sample with a grain fineness number within the target range. These results suggest that in terms of grain size the location of the Wexford Sand Company is not anomalous and sand deposits with similar grain size distributions and similar levels of variability can be found in other areas of the state. 93 Grain Fineness Number for Glacial Outwash Samples Count < 44.0 44.0 - 46.9 47.0 - 53.0 53.1 - 56.0 > 56.0 GFN Range Figure 5.17: Frequency distribution of grain fineness numbers for glacial outwash samples. is H\ l. x ,L.‘ l‘ I“. 0 30 601<1|ometers K 7 “ “ ‘L " Sample Locations . . 2:4! A inland dune _- __ _ ,. ' V ‘ . 4° '- 0 Glacial outwash - Vi 9 . k W Standard Deviation. GFN ' i . $3.00 0 . 3.01-6.00 O A 6.01-9.00 O A 901-1200 0 A >12.oo 0 Not Applicabb - Railroad Corridor [ County Boundary * Wexford Sand Co. Figure 5.18: Variability in grain size distribution based on the standard deviation of GFN for sites with samples from multiple depths. Low standard deviations are desired, as they signify more consistency of GFN at a given location. 94 As mentioned above, the grain fineness number (GF N) is the AF S statistic for summarizing grain size distributions of samples that have a wide range of grain sizes. However, just as two different sets of numbers can have the same mean, it is possible for two sand samples with varying grain size distributions to have the same grain fineness number. Therefore, it is important to also look at the percentage of sand (by weight) retained on each sieve or group of sieves during sieving. Foundry sand users have set specific target values and ranges for the percentage of sand that should be retained on certain size sieves or groups of sieves (Table 4.1). It is possible for a sample to have a GFN within the target range but not meet all acceptable percentage ranges for each sieve or group of sieves. An extreme example is Newaygo 3 (Figure 5.7b), which has a GFN of 47.6 but does not meet any of the acceptable ranges for specific sieve sizes (Table 4.1). Due to this possibility both GFN and percentage retention on specific sieves or groups of sieves are used as a means of quantifying the grain size distribution of a sample. For each sample, the percentage of sand retained on each of 8 different size sieves (or groups of sieves) was compared to the appropriate acceptable range (as outlined in Table 4.1). The total number of sieves or sieve groups on which the percentage retained fell within the acceptable range was tallied for each sample. For example, if the percentage of sand retained on the #40 sieve (420 microns) was between 4.5% and 13.0% then that would be one acceptable sieve for that sample. If the percentage of sand retained on the #50 sieve (300 microns) was 37.5% then that sieve would be unacceptable because this percentage is outside of the target range. The possible range of this statistic is from 0 to 9. 95 Actual tallies from all of the samples collected ranged from 0 to 8 (see Appendix C). No glacial outwash samples were acceptable on all sieves, including sand currently being mined for foundry use by Wexford Sand Company. Unprocessed samples fi'om Wexford Sand Company were acceptable on between 2 and 7 sieves (or sieve groups). Fourteen glacial outwash samples (10%; not including those fi'om Wexford Sand Co.) were within the acceptable range on at least 7 of the 9 categories (Figure 5.19). Thirty- nine outwash samples (28%) were within the acceptable range on at least 5 of the 9 sieves. r 1 0 30 60 Kilometers l——+—-l N A. Sample Locations ' A Inland dune 0 Glacial outwash Sieve Analysis (max 9) e A 7 or more acceptable sieves 0. ' >Kalk ka 0 A 6 or fewer acceptable sieves 9 .K - Railroad Corridor V9 1 g / ‘ K \ \ :1 County Boundary :31 b. a . \ ‘ Wexford Sand Co ord “#0 d 9‘ \\ < 1' e Newaygo} o 5 1.x . , 1W, 4 ' \ e . _ \fl \ ‘\ ‘ L] irx \ Kw \/\ .) \ _ J l D ‘1 f , I / ‘ i \\ Cl v . A T Figure 5.19: Results of the sieve analysis. Labeled locations are acceptable on 7 or more sieves. 96 Inland Dune Locations The median GFN (64.1) and the mean GFN (63.5) of inland dune samples were both above the target range. The GFN range of 38.1 to 83.1 for inland dune samples (Table 5.2) was narrower than the GFN range for glacial outwash samples. As with outwash samples, the grain size distributions of inland dune sand showed high variability. At no inland dune locations did all samples fall within the target range for GFN. Only one inland dune sample (from Chippewa 1) was within the GFN target range. By using the Mann-Whitney non-parametric statistical test, the difference in median OF N between glacial outwash samples (48.2) and inland dune samples (64.1) was shown to be significant at the 0.05 level (U = 2326, p = 0.00). The median GFN is statistically higher for inland dune sand than for glacial outwash sand, signifying a slightly finer texture for inland dune sand, which concurs with the findings of Sundeen (1978a). Standard deviations of GFNs for inland dune sand range from 1.47 to 10.69 (Table 5.2; Figure 5.18). The median standard deviation of GFN for inland dune samples was 5.61 and the mean was 5.67. The Mann-Whitney test showed at the 0.05 level that there is no significant difference between the mean standard deviation of GFN for glacial outwash samples and that of inland dune samples (U = 106, p = 0.784). Thus, while inland dune samples tended to average a slightly higher GF N than glacial outwash samples the GFN variability does not significantly differ between the two geomorphic settings. Two inland dune locations (Chippewa 4 and 5) exhibited less variability in grain size than the Wexford Sand Company sand mine. Inland dune samples met between 1 and 8 target percentage ranges of the sieve analysis. No samples met the maximum possible number of 9. One of 20 inland dune 97 samples (5%) was within the acceptable range in at least 7 of the 9 categories (Figure 5.19). Nine samples (45%) were within the acceptable range on at least 5 of the 9 categories. Table 5.2: Grain Fineness Numbers for inland dune samples. Standard deviations represent variability among multiple samples from the same location. # Slaves Standard Deviation Location Sample GFN (Target 47-53) (Max 9) of GFN Arenac 2 A 72.5 4 Arenac 2 B 66.8 5 Arenac 2 C 61.3 4 Arenac 2 D 53.4 8 8.14 Chippewa 1 A 59.4 4 Chippewa 1 B 47.1 5 Chippewa 1 C 38.1 1 10.69 Chippewa 2 A 70.5 3 Chippewa 2 B 71.0 3 Chippewa 2 C 83.1 4 Chippewa 2 D 77.6 4 5.98 Chippewa 3 A 58.3 6 Chippewa 3 B 68.7 5 Chippewa 3 C 62.3 5 5.25 Chippewa 4 A 59.9 4 Chippewa 4 B 57.4 5 Chippewa 4 C 60.0 4 1.47 Chippewa 5 A 70.5 5 Chippewa 5 B 65.9 3 Chippewa 5 C 66.7 5 2.46 98 25-micron Clay Percentage clay is an important parameter of foundry sand, as excessive clay can inhibit correct bonding during the mold and core making process. Since current molds and cores do not use natural clay as a bonding agent, clay content of ideal foundry sand should be as low as possible. According to the United States Department of Agriculture (USDA), clay is considered to be any particle with a diameter of less than 2 microns. The USDA classifies particles between 2 and 25 microns as fine silts (Schoeneberger et al., 2002). The foundry industry groups these two particle size groups together in a category it calls ‘25-micron clay’ (American Foundry Society, 2004). Results of the clay analysis confirm that many sand deposits in Michigan contain very low amounts of clay. Seventy-nine samples, from 33 locations, contained less than 1.0% 25-micron clay. At 11 locations, all samples had less than 1.0% 25-micron clay (Figure 5 .20) Glacial Outwash Locations Clay content in glacial outwash samples was minimal in most cases. Sixty-two glacial outwash samples (45%) from 27 locations contained less than 1.0% 25-micron clay (Figures 5.20 and 5.21; Appendix C). The median clay content for glacial outwash samples was 1.0% and the mean was 2.0%. Eight of nine outwash locations in Michigan’s Upper Peninsula had at least one sample with less than 1.0% clay (Figure 5.20). Because clay can be translocated by water moving through the soil it can easily become concentrated in the subsurface. Samples closer to the surface have a higher 99 K \_ . \w K Chl \j l ,2 N“.: Eli/I \. 7.] 71/ '11- ‘1 1 Sample Locations A Inland dune . 0 Glacial outwash ‘ 25-micron Clay . .3 e . e e A All samples < 1.0% clay 9 jQV/ K e A 1 or more sample 21.0% clay h .\‘ ‘ i. . b . /‘ i \\ ‘ - Railroad Corridor 1‘ -l “\ ‘ i ‘ ~\ l\ - County Boundary K’ inks 1\: I —‘ I \ * Wexford Sand Co. 1 .‘. i. l: 0‘: / 9 l . < . ‘ , L/ l > / n 7 I l / \ / / v L Figure 5.20: Clay content results for all sample locations. Labeled locations are those in which all samples had less than 1.0% clay. Glacial Outwash Samples Count 0.0-0.9 1.0-1.9 2.0-2.9 3.0-3 .9 % 25-mlcron clay 4.0+ Figure 5.21: Frequency distribution of 25-micron clay content for glacial outwash samples. 100 possibility of accumulating this translocated clay, especially given the high infiltration rates in sandy soils. Therefore, it is beneficial to compare clay content of the deepest sample at each location, since this would be expected to be least susceptible to accumulating translocated clay. At 17 outwash locations the deepest sample had less than 1.0% 25-micron clay (Figure 5.22). “ Chlp 0 30 60 Kilometers \\ \\ H\_ r c 7 F—+———OW>OCD>OOW>OW>UOW> mum-##szwwNNNN-A-A-s 103 CHEMICAL ANAL YSIS 0F SANDS Important chemical factors that determine whether a sand source can be viably used in the foundry process include pH and Acid Demand Value (ADV). These two variables affect the chemical reactivity of the sand, which must be subjected to multiple chemicals during the casting process, such as acid catalysts and binding resins. Results of the pH and ADV analyses are described in the following sections and have been grouped by geomorphic setting, either glacial outwash or inland dunes. Samples used in the chemical analyses were collected from the locations shown in Figures 5.7a, b, and c. pH The pH value of a sample gives the water soluble level of alkalinity or acidity of the sand. This determines its reactivity with resins used in the casting process. Sand with a pH that is close to neutral (7.0) is the least reactive and therefore best for casting. The target range of pH for foundry sand is 6.5 to 7.8. Because sandy deposits have high infiltration rates, slightly acidic rainfall and snowmelt leaches naturally-occurring carbonates from near-surface sediments. As a result, most sandy soils in Michigan register an acidic pH even when formed out of carbonate-rich, higher-pH parent materials (Schaetzl and Weisenbom, 2004). The pH of shallow samples may be affected by leaching and thus may not be reflective of the underlying parent material. Because of this, only the deepest sample below 3.0 m from each location was pH-tested, to most closely represent the pH of the parent material. Upper Peninsula samples exhibited generally low pH while higher values were found in the Lower Peninsula (Figure 5.23). Few locations in the Lower Peninsula had a 104 pH below 6.5. In general, samples from the Lower Peninsula were slightly alkaline while all but one sample from the Upper Peninsula was acidic. l, Sample Locations A Inland dune 0 Glacial outwash pH - deepest sample below 3.0 m e A 5.12 - 6.49 e A 6.50 - 7.80 e A 7.81 - 8.56 - Railroad Corridor [:l County Boundary * Wexford Sand Co. Figure 5.23: pH of the deepest sample taken at each location (if below 3.0 m) 105 Glacial Outwash Locations Thirty-one glacial outwash locations were tested for pH (Table 5 .4). In addition, 4 samples from Wexford Sand Co. were tested for comparative purposes (Table 5.4). The range of pH from outwash locations was 5.81 to 8.56. Median pH for glacial outwash locations was 7.51 and mean pH was 7.28. Thirteen of 31 glacial outwash samples (42%) fell within the foundry sand pH target range of 6.5 to 7.8. Only one of the samples meeting the target criteria was located in the Upper Peninsula. All other Upper Peninsula outwash samples had pH values below the target range. This may be due to the relationship between underlying bedrock type and soil pH (e.g. Searcy et al., 2003). Given the relative lack of carbonate-rich bedrock in the Upper Peninsula compared to the Lower Peninsula (Dorr and Eschman, 1970) it is not surprising that pH values of sand from the former region are generally below the target range. Results fi'om the Wexford Sand Co. mine ranged from 7.85 to 8.01 , slightly above the target range for foundry sand. Inland Dune Locations Five inland dune locations were tested for pH. The pH range for inland dune samples was 5.12 to 7.95 (Table 5.5). Median pH for inland dune locations was 6.14 and mean pH was 6.41. One of 5 inland dune locations (20%) was within the pH target range. The non-parametric Mann-Whitney test was used to test for a significant difference in pH based on geomorphic setting. The test found that the difference in pH between glacial outwash and inland dunes is not significant at the 0.05 level (U = 46.0, p = 0.15). An additional test was done to look for differences in pH between Upper 106 Table 5.4: pH and Acid Demand Value for glacial outwash locations (deepest sample below 3.0m) Location Sample pH ADV _A_lger 1 D 6.02 0.70 flger 2 C 6.30 0.90 flger 4 D 6.40 0.30 flger 5 D 6.06 0.00 Alger 6 l 7.38 1.23 Allegan 1 C 5.95 0.00 Allegan 2 D 7.44 19.51 Allegan 3 D 7.85 34.65 Antrim 2 D 7.90 35.83 Chippewa 7 D 5.82 0.72 Crawford 1 D 7.94 37.27 Crawford 2 D 7.85 45.69 Crawford 3 D 7.88 36.77 Grand Traverse 1 D 7.50 1.30 Grand Traverse 2 D 7.93 38.09 Kalkaska 1 D 7.84 37.43 Kalkaska 2 C 7.79 35.07 Lake 1 D 7.59 34.77 Lapeer 1 C 8.08 38.32 Newayg) 1 D 7.51 3.69 Newayg2 D 5.81 0.92 Newaygo 4 D 7.07 2.26 NewayggS D 7.46 6.61 _ggemaw 1 D 6.90 0.40 Otsego 1 D 7.73 5.44 Otsego 3 D 6.26 0.00 Wexford 1 D 7.70 6.91 Wexford 2 B 8.12 41.18 Wexford 5 D 7.62 8.12 Wexford 6 C 8.56 42.79 Wexford 7 D 7.30 1.60 Wexford Sand Co. 1 7.85 43.39 Wexford Sand Co. 2 7.98 36.18 Wexford Sand Co. 3 8.01 37.40 Wexford Sand Co. 4 7.88 33.33 107 Table 5.5: pH and Acid Demand Value for inland dune locations (deepest sample below 3.0 m) Location Sample pH ADV Arenac 2 D 7.95 47.01 Chippewa 1 C 6.72 36.45 Chippewa 2 D 6.14 0.00 Chippewa 3 C 5.12 0.00 Chippewa 5 C 6.10 0.00 Peninsula and Lower Peninsula samples. The Mann-Whitney test indicates a significant difference in median pH when samples are grouped by location (Upper Peninsula vs. Lower Peninsula) as opposed to geomorphic setting (U = 236.0, p = 0.00). This difference may be related to the difference in parent material in these two localities (e.g. Searcy et al., 2003). Acid Demand Value Just as pH gives a water soluble level of alkalinity or acidity of the sand, acid demand value (ADV) gives a measure of the chemical reactivity of the sand with an acid. Since acid catalysts are often used in the mold- and core-making processes, acid demand value is an important characteristic of foundry sand. High acid demand values can alter reaction speeds and strengths of sand/binder bonds. Ideal acid demand value of foundry sand is 10.00 or less. The range of AF S acid demand test possible values is from 0.00 to 50.00. Results from this research ranged from 0.00 to 47.01. Given the potential issues of leached carbonates described previously, testing for acid demand was restricted to the deepest sample at locations where samples were collected at depths greater than 3.0 m. The distribution of values is bimodal with one peak between 0 and 5 and a second peak 108 between 35 and 40 (Figure 5.24). Only one sample had an acid demand value between 10 and 30. This bimodality has a limited geographic component. All Upper Peninsula ADV values except one were less than 5.00. Lower Peninsula ADV values, however, ranged across the entire spectrum. Results suggest a moderate correlation between pH and ADV. Samples with acidic pH tended to have low ADV values and samples with basic pH largely had high ADV values. If pH was near neutral, ADV varied from low to high. The reason for a general lack of mid-range values is unknown, but it may imply that the presence or absence of some substance, such as carbonates, plays an important role in regulating ADV. Median acid demand value is 6.03, evidence of the relatively large number of samples with low ADV. Frequency Figure 5.24: Frequency distribution of Acid Demand Value (ADV). Ideal values are below 10.00. 109 Glacial Outwash Locations Acid demand value (ADV) of glacial outwash samples ranged from 0.00 to 45.69, with a median of 6.61 and a mean of 16.72 (Table 5.4). Given the bimodal distribution of glacial outwash ADV, the mean value is not representative of the underlying distribution. Eighteen of thirty—one glacial outwash locations (58%) met the foundry sand criteria for ADV. Six of these locations were from the Upper Peninsula whereas twelve were from the Lower Peninsula (Figure 5.25). Sand currently mined from Wexford Sand Company’s inland site was also tested for ADV. Unprocessed sand collected from multiple locations at the site had acid demand values that ranged from 33.33 to 43.39. Such sand must be processed to make it ready for foundry use. Nineteen glacial outwash locations in this study have lower acid demand values than all samples taken at Wexford Sand Company’s inland site, while 30 locations have acid demand values that are lower than at least one of the samples collected from this sand quarry. Wexford Sand Company’s use of such sand suggests it is possible to preprocess sand in order to artificially lower acid demand value and still remain economically viable. Inland Dune Locations Acid demand value at inland dune locations ranged from 0.00 to 47.01, with a median of 0.00 and a mean of 16.69 (Figure 5.25; Table 5.5). Three of five inland dune locations (60%) analyzed had acid demand values of less than 10.00, meeting foundry sand criteria. Because the median and mean values for ADV are not representative of the underlying bimodal distribution a statistical comparison of means or medians based on 110 geomorphic setting (glacial outwash or dunes) or location (Upper Peninsula or Lower Peninsula) was not deemed appropriate. Even so, a nonstatistical comparison of the data seems to show that for acid demand value, like pH, location (Upper Peninsula vs. Lower Peninsula), and not geomorphic setting, appears to play a more significant determining role. 0 30 60 Kilometers I—i—l \“x Sample Locations A inland dune 0 Glacial outwash Acid Demand Value e A 0.00-5.00 e A 501-10.00 e A 10.01-40.00 e A 40.01-47.01 - Railroad Corridor _ County Boundary iv Wexford Sand Co. Figure 5.25: Acid demand value (ADV) for all tested locations. Locations with values less than 10.00 are best suited for foundry sand and have been labeled. 111 OVERALL SUITABILITY This chapter presented the results of economic, physical, and chemical analyses of glacial outwash sands and inland dune sands fi'om sites that lie within 8 km (5 miles) of the railway corridors in Michigan. These results indicate that the amount of potential sand for foundry use in inland Michigan dwarfs the sand resources of coastal dune areas. In addition, almost one-third of the state’s land area is within 8 km (5 miles) of an existing railroad line. This minimizes the need to build costly new transportation networks to access and extract inland sand. Railroads can be used in conjunction with barge transportation to access inland sand resources in Michigan’s Upper Peninsula. The thickest deposits of inland sand occur in northwest Lower Michigan and are glacial outwash sand. Deposits in this region can be over 75 m thick and are a legacy of the Wisconsin glaciation (Rieck and Winters, 1993). Estimated sand thickness is at least 20 m across almost all of northern Lower Michigan, except the extreme northeast, where bedrock is quite shallow. The most widespread near-surface deposits of inland sand are found in north-central Lower Michigan (Figure 5.3). An approximately 50 km by 50 km area covering parts of Antrim, Otsego, Kalkaska, and Crawford Counties is estimated to have 20 or more meters of sand within 23 m of the surface. This translates to approximately 50 cubic kilometers of near-surface sand in this region of the state alone. Physical testing suggests that most inland sand deposits in Michigan have grain shapes ideal for use by the foundry industry. Almost all samples had the requisite AF S subangular grain shape. The only locations at which all samples did not meet grain shape criteria were Allegan 6, Chippewa 3 and 6, Tuscola 1, and Wexford l, 2, and 4. 112 Results for grain size were quite variable. At no location did every sample meet all of the grain size criteria (GFN and sieve percentage ranges). However, many locations contained sand that was as good or better in terms of grain size than sand currently mined from the Wexford Sand Co. mine. Processing of sand to create different grain size mixtures from the same deposit is common practice in sand mining and most locations in this analysis contain grain size distributions that are reasonably close to the desired distribution. Clay content was more consistent than grain size and was generally very low at most locations. Many samples contained very little 25-micron clay. A statistically significant difference was found between Upper Peninsula and Lower Peninsula samples, with the former averaging less clay than the latter. Chemical analysis indicates multiple locations that meet both pH and acid demand value (ADV) criteria. Many of these locations are in the northwest Lower Peninsula, where sand thickness is also greatest. ADVs of all but one Upper Peninsula sample met the target criteria; however, all but one of these samples had pH values that were below the target range. The effects of deep leaching of carbonates on pH and ADV results, while minimized, may still be present. Statistical tests found no significant differences for median pH based on geomorphic setting. However, differences between Upper Peninsula and Lower Peninsula samples were found to be significant. Both mean pH and mean ADV were generally lower for Upper Peninsula samples than for Lower Peninsula samples. By combining the results of all 7 physical and chemical tests it is possible to rank sample locations based on overall suitability. As stated previously, no locations, 113 including the sand mine operated by Wexford Sand Co., fell completely within the bounds of suitability on all 7 tests. The results of each test were scored on a scale of 0 to 4 for each location, with a score of 4 meaning that the location met all foundry parameters for that test (Tables 5.6 and 5.7; Appendix D). A combined score of 28 would signify a location at which all variables tested were found suitable for all samples. Of the locations at which all 7 variables were measured, the best results were in northern Lower Michigan (Table 5.6, Figure 5.7b). Three locations in Wexford County (Wexford 1, 5, and 7), one in Grand Traverse County (Grand Traverse 1) and four in Newaygo County (Newaygo 1, 2, 4, and 5) scored very high and are in areas estimated to have thick and extensive near-surface sand deposits. Upper Peninsula locations with high suitability are mostly in Alger County (Table 5.6). Due to the depth restriction imposed on testing pH and ADV and the need for multiple samples from the same location to calculate standard deviation of GFN, not all locations were tested for all 7 variables. Suitability results based on the 4 variables tested at almost every location (grain shape, GFN, sieve percentages, and 25-micron clay) are listed in the “Subtotal” column of Tables 5.6 and 5.7. Overall, the most likely candidates for replacing coastal dune sand are in three distinct zones: Wexford and southeast Grand Traverse Counties, central Alger County, and northern Newaygo and southern Lake Counties (Figure 5.26a, b). Not all targets were met with all samples tested but combined results suggest the amount of processing necessary to prepare sand from these zones for foundry use (in regards to those characteristics tested) would be lowest among deposits that were sampled in this study. 114 mF FF o o N N F. n F. N mass: 220 FN 2 v F. m N n F F. F 8.93 290 -- m 1 1 e F F. o F. F. 22390 NF m o o n F n F 4 m 2°38 FF m o o F F N a F. N 90390 NF a o o n F N F F. F 2°?an NF m a. o N N o o v N mzwmmzo -- N 1 1 F F m o o m g 11 m | | m N l O v r omcwh< 2 NF 0 o m N F. n F. N E_=F_< -- s 1 1 o 1 1 1 v F 55,2 1 N 1 1 m N o o N e 6&2 1 .2 1 1 m N 1 v F. m .532. 1 m 1 1 m N F. o v v 5mm? m N o o o N N F v m :32? E N o F. F F n F v N 59.2 2 NF F. o m N N n v F 58.2 2 s F. F. N F F F v m am? 2 NF F. o m N m N v m am? 2 m w o N F v o F. F. ma -- m 1 1 m N F o F. N 5% FF a v o N F o o v N am: NF NF F. o m N F n F. F mag Eon 5°35 >o< : as 8.25.3 855 8350 2.6 «mam Salas... 3853 20.0 520 $29 5.33.13 ES £0.57. .ZmU .0927. £9.» .8 3...: be 5.3. 2: 3 55:3 133.13. 2; .ESoEfl—eg bee—.8 5.2.3 0.8 833...; =a 83%.: 2.5.: an .«e 0.58 FEE—:3... < 6.5332 :mafiae Eon—u =a .5.— mueieaa bin—azam "9m 935—. 115 NF NF o o N N N N v 3 NF N F. F. N N N o F. N 2255 N N o o o F N o F. N 9262, NN NF F. F. N N N F N N News; 1 N 1 1 o F o o N F. sons; 1 N -- 1 o N 1 N N N 92me N F. o o o N F o N N Eons; NN NF F. F. N N e F. N F 226; I c 1 .. o 1 l I 1 F 208:... 1 N 1 1 N F 1 o F. F 5.28:8 NF NF F. o N F N N F. N cameo 1 N 1 1 F F 1 o F. #81201 NF N F. F. N F N o F. F cameo 1 FF 1 1 F N 1 F. v N 32:80 FN FF 4 N N N N N F. F 25580 FN NF F. F. N N N F F. N omxssmz NN N v v N N N F N F. omxcsmz 1 NF 1 1 N o 1 F. F. N omxssmz NF NF F. o N N N N F. N omxssmz NN N F. F. N F F. F N F omNNzaz 1 N 1 1 o F 1 o F. Nam N N o o F F o o F. [memfll NF N o F. N F v o F. F 9.3 NF NF 0 F. N N N F F. N 9.8.5. N N o o N F F o F. F 9.853. .32 333% >o< : so SEENN SEN Season 26 new 5:83 tun—9:5” E Eaumv dress N.N asap 116 NF N F. o N N F. o F. N 3895 1 N 1 1 N N F. o F. F. 3895 NF N F. o N N N o N N $895 NF NF F. o F. N N o F. N $895 NF NF 0 F. N N F F F. F 36820 FF N o o N N N o F. N 852 F38 5895 >3 1.. Eu :22... NN 8.55 5.638 :3 82m 5:83 2.2.5». 2.6 526 nae—o conga—1mm use .853 .ZRU .0923 5a..» Lou Nos—a.» No 5.3. 2: N. 55:8 .3633 2:. £805.22— EFF—Sou £5?» «.3 83st? =a 33%.: 2:95 mu no 0.5% 55:53:. < .2533... 2:... tea—5 =a Lou Nana—ea.— bzfiazam "Nam 03:5. 117 Suitability Scores O 21 - 24 C 17 - 20 O 13 - 16 O <13 0 no pH or ADV Most Suitable Areas m Primary 7//A Secondary Estimated Cumulative Sand Thickness (m) Upper 23m [:1 Rail Corridor Figure 5.26a: Primary and secondary suitability zones in northwest Lower Michigan. Boundaries are generalized based on railroad corridors and estimated sand thickness. 118 Suitability Scores 0 21 - 24 O 17 - 20 O 13 - 16 e <13 0 no pH or ADV Moat Suitable Areas m Primary 7//A Secondary Estimated Cumulative Sand Thickness (m) Upper23m -o-5 -5-10 -10-15 -15-20 -20+ D Rail Corridor - ‘ Figure 5.26b: Primary and secondary suitability zones in the Upper Peninsula. Boundaries are generalized based on railroad corridors and estimated sand thickness. 119 Each of the three primary zones listed tested well for most variables but weaker for some. Highly suitable locations in Alger County generally had high rankings for grain shape, grain size, and clay content, but pH was below the optimum range. In Wexford County, some locations scored very well, while others scored quite low, suggesting sand deposits in this area may be quite variable and further testing would be prudent. Highly suitable locations in Newaygo and Lake Counties generally scored well for pH and ADV but lower for clay content and GFN. All things considered, these three zones would likely require the least amount of processing to be used by the foundry industry and have sand that is just as good as or better than sand currently being mined from the Wexford Sand Co. mine. Each foundry places slightly different importance on the characteristics tested here, based on their own individual needs. By listing the many benefits and few less suitable characteristics of each of the primary zones, individual foundries can better focus on the area that best meets their needs. Secondary areas that may be suitable but would likely need additional testing include the Trout Lake area of Chippewa County as well as a larger area covering portions of northern Grand Traverse, Antrim, Kalkaska, Otsego, and Crawford Counties (Figures 5.26a, b). The latter group of five counties did not score as well as the primary zones described above, however, sand volume in this region is immense and further testing may reveal subsections that rival or exceed the primary zones in suitability. The Trout Lake area of Chippewa County (Chippewa 1 through 5; Figure 5.26b; Table 5.7) shows moderate suitability but is somewhat confined in extent. Further study would be needed to refine sand volume estimates in this area and to determine whether the 120 inland dunes here could support a fiill-scale sand mine. Estimated sand thickness in this area probably includes lacustrine and/or outwash sands underlying the dunes, which would likely have characteristics that differ from the eolian sand above. Suitability may also be high around the location of Ogemaw 1, but given the limited number of samples in this area any recommendation should be tempered. In general, glacial outwash samples showed a slightly higher degree of suitability than inland dune samples for foundry use (Tables 5.6 and 5.7). Statistically, glacial outwash sand and inland dune sand exhibited differences in grain size distribution and clay content. Grain shapes of both groups were similar. Mean and median GF N were higher for inland dune sand than for glacial outwash sand, indicating that inland dune sand was generally finer in texture. Mean and median GFN were within the target range for glacial outwash sand but above this range for inland dune sand. GFN variability (of multiple samples from the same location) showed no statistical difference between outwash and inland dunes. Median clay content was statistically lower in inland dune samples. Even though glacial outwash sand proved only slightly better than inland dune sand in terms of physical and chemical suitability, the volume of sand contained in inland dunes is dwarfed by that of outwash deposits. Because of this, few inland dune locations were initially selected as sampling locations. Estimated sand thickness at these inland dune locations probably reflects underlying outwash or lacustrine sands, since the inland dunes themselves are generally quite small, especially compared to their coastal counterparts. Inland dunes are also generally scattered across the landscape and not 121 continuous like outwash deposits. Economically, glacial outwash deposits are far more suitable for foundry sand use than inland dunes because of these differences. SUMMARY Within the primary zones of suitability, extensive sand deposits have been found; tests have confirmed the presence of sand that is low in clay, has a suitable or near- suitable range in grain size, has the desired subangular grain shape, and has pH and acid demand values that are close to industry targets. A well-sited inland sand mine in one of these regions could potentially have enough sand resources to provide for decades of operation. Combined results of the economic, physical, and chemical tests have shown that areas of both northern Lower Michigan and the Upper Peninsula have the potential to replace current coastal dune sand mines. 122 Chapter 6 CONCLUSIONS AND RECOMMENDATIONS The goal of this research was to determine whether inland sand deposits in Michigan have the potential to replace coastal dune sand in foundry molds and cores. A GIS was used to determine the best areas for sampling inland sand based on accessibility and estimated sand thickness. Within these areas, 53 locations were selected and sampled. At each location, up to 9 samples, but typically 4, were tested for grain shape, grain size distribution, 25-micron clay, pH, and acid demand value. Results from these tests were compared to general foundry sand requirements. In addition, they were ranked against sand collected from a sand mine in Wexford County that has supplied foundry sand to Ford Motor Company. Statistics were calculated to check for significant differences when results were grouped by general geomorphic setting (glacial outwash or inland dunes) and/or location (Upper Peninsula or Lower Peninsula). In terms of accessibility, Michigan’s inland sand is highly satisfactory. Over one- third of the state is within 8 km (5 miles) of a railroad. Within these railroad corridors exist large volumes of sand. Estimates of sand thickness suggest the largest amounts of inland sand are concentrated in northern Lower Michigan. In this same area, the estimated amount is likely an underestimate of the true amount of sand. Physical and chemical tests on sand collected fiom within these railroad corridors suggest that it has most of the characteristics required to be used by the foundry industry. Grain shape is satisfactory at almost all locations. Acid demand value and pH are satisfactory at many locations and 25-micron clay content is largely minimal. The grain 123 size distribution at any given site can be quite variable, however, which means that inland sand would almost certainly need to be artificially sorted prior to foundry use. Such artificial sorting, though, is not uncommon, especially since sand mine operators often market to a variety of users. When ranked by combined test results 3 regions show the most potential to be inland sand sources for the foundry industry: 1) Wexford and southeastern Grand Traverse Counties, 2) northern Newaygo and southern Lake Counties, and 3) central Alger County. These areas tested highly in both physical and chemical sand characteristics. They are within 8 km (5 miles) of existing rail lines and are estimated to have large volumes of near-surface sand. Secondary inland areas that also may have the potential to replace coastal dune sand include: 1) eastern Antrim and western Otsego Counties (and into northern Grand Traverse, Kalkaska, and Crawford Counties), and 2) the Trout Lake area of southwest Chippewa County. These areas are also located close to existing transportation networks and generally have very thick sand deposits. However, they tend to be either more limited in extent (the Trout Lake area) or less amenable physically and chemically than the primary regions. Foundry sand from these areas would likely require more preprocessing than sand from the primary regions listed above. The conclusions of this research have considerable implications for foundry sand mining in Michigan. Large amounts of potential foundry sand are located in accessible inland areas. Given the political, environmental, and ethical controversy surrounding the mining of coastal dunes, a move to inland sources is clearly possible. Such a move would not be free from its own controversy, but should be explored as a better and more 124 ethical means of balancing the economic, environmental, and recreational value of Michigan’s natural resources. 1 Current environmental regulations magnify the benefits of inland sand relative to coastal dune sand. While inland sand deposits are also a component of ecological systems, these systems are generally not regarded as being at the same level of risk as Michigan’s coastal dune environments. As such, regulations are less restrictive in inland areas. This has the potential to offset any decrease in the physical and chemical quality of inland sand relative to coastal dune deposits. Foundry sand users and sand mine operators may wish to apply these new data to their specific requirements. Because mold- and core-making processes are not identical at every foundry, physical and chemical sand specifications can vary to a limited degree. Individual users may find inland sand somewhat more or less amenable to their own specific needs. In addition, non-foundry uses of the vast inland sand resources can be explored. Because many non-foundry sand users have less stringent physical and chemical requirements than foundries, the potential available amount of inland sand is even greater. This study also highlights the need for continued investigation of Michigan’s inland sand resources. Over the past 30 years there has been a dearth of published studies on Michigan’s industrial sand resources, even though such studies have been both needed and explicitly requested. Future studies should address the following important areas: 1) Detailed investigation of the primary inland sand areas identified by this research, including their extent and the variability within them. 125 2) The feasibility of processing sand in secondary areas for foundry use. 3) The economic variables, beyond distance and mode of transportation, that affect market access to sand resources. 4) Non-foundry uses for which sand with these physical and chemical specifications would be appropriate. 5) The potential of sand from adjacent states to replace Michigan coastal dune sand in foundries. Based on the results of this research, a move to inland sand sources and away from coastal dune mining is feasible. Such a move has the potential to benefit the interests of many vested parties, including sand mine operators, recreational dune users, and all those who place intangible value on coastal dune ecosystems. Further study can refine these results and lead to even more specific recommendations and direction toward inland sand resources. 126 Appendix A Table A1: Sample locations. Location Latitude Longitude Alger 1 46° 20' 55" 86° 34' 43" _A_I_ger 2 46° 20' 15" 86° 30' 43" flger 3 46° 22' 55" 86° 36' 22" flger 4 46° 21' 23" 86° 44' 07" Alger 5 46° 19' 13" 86° 42' 41" flger 6 46° 23' 12" 86° 39' 19" Allegan 1 42° 39' 25" 85° 59' 38" Allegan 2 42° 39' 03" 86° 00' 56" Allegan 3 42° 36' 31" 85° 58' 16" Allegan 4 42° 29' 33" 86° 02' 46" Allegan 5 42° 32' 59" 86° 03' 24" Allegan 6 42° 30' 54" 86° 04' 20" Antrim 1 44° 56' 55" 84° 56' 48" Antrim 2 45° 01' 23" 84° 58' 03" Arenac 1 44° 04' 21" 84° 03' 24" Arenac 2 44° 04' 19" 84° 03' 22" Chippewa 1 46° 13' 12" 85° 01' 15" Chippewa 2 46° 11' 49" 85° 01' 24" Chippewa 3 46° 11' 40" 85° 02' 39" Chippewa 4 46° 10' 02" 85° 02' 20" Chippewa 5 46° 14' 44" 85° 00 36" Chippewa 6 46° 16' 27" 84° 26' 10" Chippewa 7 46° 17' 37" 84° 35' 31" Crawford 1 44° 46’ 32" 84° 44' 28" Crawford 2 44° 38' 10" 84° 40' 05" Crawford 3 44° 34' 58" 84° 39' 22" Crawford 4 44° 35' 49" 84° 42' 17" Grand Traverse 1 44° 32' 29" 85° 24' 16" Grand Traverse 2 44° 43' 34" 85° 22' 56" Kalkaska 1 44° 47' 11" 85° 07' 03" Kalkaska 2 44° 48' 51" 85° 08' 00" Lake 1 43° 53' 08" 85° 57' 29" Lapeer 1 43° 06' 43" 83° 14' 04" Lapeer 2 43° 07' 27" 83° 16' 28" Newaygo 1 43° 28' 47" 85° 46' 57" Newaygo 2 43° 35' 45" 85° 50' 13" Newaygo 3 43° 23' 44" 85° 43' 29" Newaygo 4 43° 42' 54" 85° 48' 00" Newaygo 5 43° 47' 17" 85° 53' 24" Qgemaw 1 44° 19' 42" 84° 13' 03" O emaw 2 44° 21' 27" 84° 17' 16" Otsego 1 45° 02' 30" 84° 45' 01" Otsego 2 44° 54' 27" 84° 39' 09" Otsego 3 44° 54' 19" 84° 42' 12" Schoolcraft 1 46° 19' 50" 86° 32' 17" 127 Table A1 (cont’d). Location Latitude Lon itude Tuscola 1 43° 25' 25" 83° 29' 32" Wexford 1 44° 17' 38" 85° 38' 08" Wexford 2 44° 20' 41" 85° 31' 30" Wexford 3 44° 14' 10" 85° 20' 01" Wexford 4 44° 17' 40" 85° 21' 34" Wexford 5 44° 20' 31" 85° 28' 45" Wexford 6 44° 25' 28" 85° 21' 20" Wexford 7 44° 28' 07" 85° 23' 11" 128 Appendix B Table Bl: Sample depths. Location A QQQODODODODODCDOIUIUIOIJI#«h-hwwNNN-b-A-I A B C D A B C A B A B C D A B C D A B C D E F G H l A B C A B C D A B C D A B A Uinb¥WUQOONNNNAAA 129 Table Bl (cont’d). Location Sample Depth (m) Allegan 6 A 1.50 Allgan 6 B 3.00 Antrim 1 S 0.30 Antrim 2 A 1.50 Antrim 2 B 3.00 Antrim 2 C 4.50 Antrim 2 D 6.00 Arenac 1 A 0.88 Arenac 2 A 1.50 Arenac 2 B 3.00 Arenac 2 C 4.50 Arenac 2 D 6.00 Chippewa 1 A 1.50 Chippewa 1 B 3.00 Chippewa 1 C 3.32 Chippewa 2 A 1.00 Chippewa 2 B 2.00 Chippewa 2 C 3.70 Chippewa 2 D 7.10 Chippewa 3 A 1.00 Chippewa 3 B 2.50 Chippewa 3 C 5.90 Chippewa 4 A 0.85 Chippewa 4 B 1.70 Chippewa 4 C 2.55 Chippewa 5 A 1.50 Chippewa 5 B 3.00 Chippewa 5 C 4.50 Chippewa 6 A 1.50 Chippewa 6 B 1.83 Chippewa 7 A 1.50 Chippewa 7 B 3.00 Chippewa 7 C 4.50 Chippewa 7 D 5.26 Crawford 1 A 1.50 Crawford 1 B 3.00 Crawford 1 C 4.50 Crawford 1 D 5.22 Crawford 2 A 1.50 Crawford 2 B 3.00 Crawford 2 C 4.50 Crawford 2 D 6.00 130 Table Bl (cont’d). Location Sample Depth (m) Crawford 3 A 1.50 Crawford 3 B 3.00 Crawford 3 C 4.50 Crawford 3 D 6.00 Crawford 4 A 0.85 Crawford 4 B 1.70 Crawford 4 C 2.55 Grand Traverse 1 A 1.50 Grand Traverse 1 B 3.00 Grand Traverse 1 C 4.50 Grand Traverse 1 D 6.00 Grand Traverse 2 A 1.50 Grand Traverse 2 B 3.00 Grand Traverse 2 C 4.50 Grand Traverse 2 D 5.35 Kalkaska 1 A 1.50 Kalkaska 1 B 3.00 Kalkaska 1 C 4.50 Kalkaska 1 D 6.00 Kalkaska 2 A 1.50 Kalkaska 2 B 3.00 Kalkaska 2 C 3.54 Lake 1 A 1.50 Lake 1 B 3.00 Lake 1 C 4.50 Lake 1 D 5.50 Lapeer 1 A 1.50 Lapeer 1 B 3.00 Lapeer 1 C 4.50 Lapeer 2 A 1.50 Newaypo 1 A 1.50 Newaygo 1 B 3.00 Newmo 1 C 4.50 Newaygo 1 D 6.00 Newmo 2 A 1.50 Newaygo 2 B 3.00 . Newaygo 2 C 4.50 Newaygo 2 D 6.00 Newaygo 3 A 0.85 Newaygo 4 A 1.50 Newaygo 4 B 3.00 Newaypo 4 C 4.50 ‘ Newaygo 4 D 5.00 131 Table B1 (cont’d). Location Sample Depth (m) Newaygo 5 A 1.50 Newaygo 5 B 3.00 Newaygo 5 C 4.50 Newpaygo 5 D 6.00 _O_gemaw 1 A 1.70 Qgemaw 1 B 3.40 _ggemaw 1 C 6.80 fiemaw 1 D 11.20 ggemaw 2 A 0.77 Otsego 1 A 1.50 Otsego 1 B 3.00 Otsego 1 C 4.50 Otsm1 D 6.00 Otsgg 2 A 1.17 Otsemr 3 A 1.50 Otsego 3 B 3.00 Otsego 3 C 4.50 Otsego 3 D 4.75 Schoolcraft 1 A 1.50 Tuscola 1 A 1.50 Tuscola 1 B 1.90 Wexford 1 A 1.50 Wexford 1 B 3.00 Wexford 1 C 4.50 Wexford 1 D 6.00 Wexford 2 A 2.80 Wexford 2 B 6.00 Wexford 3 A 0.65 Wexford 4 A 1.50 Wexford 4 B 3.00 Wexford 5 A 1.50 Wexford 5 B 3.00 Wexford 5 C 4.50 Wexford 5 D 6.00 Wexford 6 A 1.50 Wexford 6 B 3.20 Wexford 6 C 4.90 Wexford 7 A 1.50 Wexford 7 B 3.00 Wexford 7 C 4.50 Wexford 7 D 6.00 132 Appendix C Table Cl: Conventional and AFS grain shapes for glacial outwash samples. Primary roundness and sphericity are the first values listed for each sample. Location Sample Conventional Grain AFS Roundness Grain 2.5 1.5 2.5 1.5 2.5 1.5 2.5 1.5 3.5 3.5 3.5 1.5 2.5 3.5 3.5 2.5 1.5 2.5 1.5 1.5 2.5 1.5 2.5 2.5 2.5 2.5 2.5 GOGQOJOQODODOIUIUIUIhhbhwwNNN—l—b—l-b A B C D A B C A B A B C D A B C D A B C D E F G H I A B C A B C D 133 Table Cl (cont’d). Location Sample Conventional Grain Roundness 3.5 3.5 2.5 2.5 3.5 2.5 2.5 2.5 2.5 2.5 xxx 3.5 2.5 1.5 2.5 2.5 2.5 1.5 1.5 2.5 2.5 3.5 2.5 1.5 1.5 1.5 2.5 1.5 1.5 2.5 2.5 3.5 2.5 2.5 2.5 3.5 2.5 2.5 3.5 2.5 3.5 2.5 Claim-54500000000 Antrim 1 Antrim 2 Antrim 2 Antrim 2 Antrim 2 Arenac 1 Crawford 1 Crawford 1 Crawford 1 Crawford 1 Crawford 2 Crawford 2 Crawford 2 Crawford 2 Crawford 3 Crawford 3 Crawford 3 Crawford 3 Crawford 4 Crawford 4 Crawford 4 Grand Traverse 1 Grand Traverse 1 Grand Traverse 1 Grand Traverse 1 A B C D A B A A B S A B C D A A B A B C D A B C D A B C D A B C D A B C A B C D 134 Table Cl (cont’d). Location Sample Conventional Grain Roundness 3.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 3.5 2.5 3.5 2.5 3.5 3.5 2.5 2.5 2.5 2.5 2.5 2.5 3.5 2.5 3.5 4.5 2.5 3.5 2.5 3.5 2.5 3.5 2.5 2.5 3.5 2.5 2.5 Grand Traverse 2 Grand Traverse 2 Grand Traverse 2 Grand Traverse 2 Kalkaska 1 Kalkaska 1 Kalkaska 1 Kalkaska 1 Kalkaska 2 Kalkaska 2 Kalkaska 2 Lake 1 Lake 1 Lake 1 Lake 1 >000!>000)>UOlIl>>UOU!>UOCD>>OCD>UOW>OW>UOCD>UOCD> 1 1 1 1 2 2 2 2 3 4 4 4 4 5 5 5 5 1 1 1 1 2 135 Table Cl (cont’d). Location Sample Conventional Grain Shape AFS Roundness Sghericlty Grain Shape Otsego 1 A 2.5 1.5. 3.5 subanfiilar Otsego 1 B 1.5 3.5 subangular Otseg; 1 C 2.5, 1.5 1.5, 3.5 subangular Otsego 1 D 2.5 1.5, 2.5, 3.5 subangular Otsego 2 A 2.5 1.5 subflqular Otsego 3 A 2.5 1.5, 2.5, 3.5 subangular Otsego 3 B 2.5, 3.5 1.5 subangular Otsego 3 C 2.5 1.5, 3.5 subangular Otsefl3 D 2.5 1 .5 subangular Schoolcraft 1 A 2.5 1.5 subanmflar Tuscola 1 A m )00: xxx Wexford 1 A 2.5 1.5 subanglar Wexford 1 B 2.5 1.5 subamlar Wexford 1 C 2.5 1.5 subangular Wexford 1 D 2.5 1.5 compound Wexford 2 A 2.5 1.5, 2.5 subangular Wexford 2 B m )00( xxx Wexford 3 A 3.5 1.5 subanguilar Wexford 4 A xxx m compound Wexford 4 B 3.5, 2.5 1.5 suba_ngular Wexford 5 A 2.5 1.5, 2.5, 3.5 subangular Wexford 5 B 2.5 1.5, 2.5, 3.5 subangular Wexford 5 C 1.5, 2.5 1.5, 2.5, 3.5 subangular Wexford 5 D 1.5 1.5, 2.5, 3.5 subangular Wexford 6 A 2.5, 3.5 1.5 subanfilar Wexford 6 B 2.5, 3.5 1.5 subangular Wexford 6 C 3.5 1.5 subagqular Wexford 7 A 2.5 1.5 subangular Wexford 7 B 2.5 1.5 subamlar Wexford 7 C 2.5 1.5 subangular Wexford 7 D 2.5 1.5 subangular ‘ 136 Table C2: Grain size distribution statistics for glacial outwash samples. Standard deviations represent variability among multiple samples from the same location. # Slevea Standard Location Deviation of GFN —L GOODCDOJCDODOJOJUIUIUIUI-b‘h-h#QQNNNA-h—I A B C D A B C A B A B C D A B C D A B C D E F G H l A B C A B C D A B C D A B wwODOOODU'l-FN-tN-IU'IhQ’v‘Ca-DNNNOD-KOOUIVUIOI-KODUINNODANN-hwmm #hwaWNNNN-aa.‘ 137 Table C2 (cont’d). # Slaves Standard Location Sample GFN (T arget 47-53) (Max 9) Deviation of GFN Allegan 5 A 50.6 5 NIA Allegan 6 A 68.6 3 Allegan 6 B 126.6 3 41.01 Antrim 1 8 xxx x NIA Antrim 2 A 48.2 4 Antrim 2 B 48.5 4 Antrim 2 C 43.8 3 Antrim 2 D 43.2 3 2.81 Arenac 1 A 57 5 NIA Chippewa 6 A 171.5 2 Chippewa 6 B 166.8 2 3.32 Chippewa 7 A 70.1 4 Chippewa 7 B 66 4 Chippewa 7 C 92.7 2 Chippewa 7 D 137.1 4 32.60 Crawford 1 A 43.4 1 Crawford 1 B 40.2 1 Crawford 1 C 48.3 5 Crawford 1 D 40.3 1 3.80 Crawford 2 A 48.9 2 Crawford 2 B 38.2 2 Crawford 2 C 49.8 5 Crawford 2 D 39.3 2 6.15 Crawford 3 A 40.6 1 Crawford 3 B 41.9 3 Crawford 3 C 50.6 6 Crawford 3 D 40.3 1 4.88 Crawford 4 A 42.4 2 Crawford 4 B 40 1 Crawford 4 C 40.6 1 1.25 Grand Traverse 1 A 42.6 1 Grand Traverse 1 B 51.3 8 Grand Traverse 1 C 41.1 1 Grand Traverse 1 D 46.8 4 4.59 Grand Traverse 2 A 44.9 4 Grand Traverse 2 B 45.7 2 Grand Traverse 2 C 47.8 6 Grand Traverse 2 D 49.2 6 1.96 Kalkaska 1 A 58.2 3 Kalkaska 1 B 41.9 1 Kalkaska 1 C 38.6 3 Kalkaska 1 D 40.2 2 9.08 138 Table C2 (cont’d). fl Slaves Standard Location Sample GFN (T arget 47-53) (Max 9) Deviation of GFN Kalkaska 2 A 45.1 3 Kalkaska 2 B 50.2 7 Kalkaska 2 C 37.3 2 6.50 Lake 1 A 40.6 1 Lake 1 B 38.2 1 Lake 1 C 43.2 1 Lake 1 D 39.7 2 2.10 Lapeer 1 A 43.2 2 Lapeer 1 B 61.3 3 Lapeer 1 C 79.2 3 18.00 Lapeer 2 A 40.9 2 NIA Newam 1 A 41.5 1 Newaygo 1 B 47.1 3 Newaygo 1 C 44.5 2 Newaygo 1 D 44.2 4 2.29 Newaygo 2 A 63.2 3 Newaygo 2 B 48.7 7 Newaxgo 2 C 61.8 4 Newaygo 2 D 51.2 7 7.34 Newaygo 3 A 47.6 0 N/A ‘ Newaygo 4 A 52 7 Newaygo 4 B 57.6 4 Newaygo 4 C 53.9 4 Newaygo 4 D 64.9 3 5.70 Newa o 5 A 52.2 7 Newa o 5 B 53.8 6 Newaygo 5 C 45.4 3 NewayQS D 41.3 2 5.85 Qgemaw 1 A 47.6 4 _O_gemaw 1 B 52.2 6 _9_gemaw 1 C 52.7 7 _Qgemaw 1 D 65.1 4 7.49 _Ogemaw 2 A 47.8 4 N/A Otsego 1 A 41 1 Otsego 1 B 44.3 4 Otsego 1 C 53.1 5 Otsego 1 D 39.1 1 6.20 Otsego 2 A 38.4 2 N/A Otsego 3 A 48.6 3 Otsefl 3 B 48.2 4 Otsefl3 C 41.8 1 ‘ Otsego 3 D 43 1 3.50 139 Table C2 (cont’d). # Slaves Standard Location Sample GFN (Target 47-53) (Max 9) Deviation of GFN Schoolcraft 1 A 36 2 NIA Tuscola 1 A m x NIA Wexford 1 A 52.3 7 Wexford 1 B 51.2 8 Wexford 1 C 48.7 7 Wexford 1 D 52 6 1.63 Wexford 2 A 59.3 5 Wexford 2 B 73.3 4 9.90 Wexford 3 A 54.9 3 N/A Wexford 4 A 60.5 2 Wexford 4 B 36.8 2 16.76 Wexford 5 A 63.7 2 Wexford 5 B 52.5 7 Wexford 5 C 57 5 Wexford 5 D 45.8 2 7.54 Wexford 6 A 32.2 2 Wexford 6 B 35.5 2 Wexford 6 C 39.7 0 3.76 Wexford 7 A 56.7 5 Wexford 7 B 45.4 3 Wexford 7 C 40.4 2 Wexford 7 D 37.6 2 8.43 Wexford Sand Co. 1 49.1 7 Wexford Sand Co. 2 48.6 7 Wexford Sand Co. 3 49.4 5 Wexford Sand Co. 4 41.4 1 3.83 140 Table C3: Percentage 25—micron clay for glacial outwash samples. Location % 25-Mlcron 0.0 0.0 0.0 0.5 1.0 0.5 4.5 0.5 0.5 1.5 1.0 0.0 0.0 0.5 0.5 0.0 0.0 0.0 0.5 0.5 0.0 0.0 1.0 0.5 0.5 2.0 0.5 0.5 0.5 2.5 2.0 1.5 2.5 7.5 2.0 1.5 3.0 1.0 1.0 1.0 1.0 1.0 6.5 OOJGQOOJODODOJUIU'IUIOI##A&Q&NNN-¥ddd Gimm$bwwwwNNNN-t—t—b A B C D A B C A B A B C D A B C D A B C D E F G H l A B C A B C D A B C D A B A A B S 141 Table C3 (cont’d). Location Sample % 25-Mlcron Clay Antrim 2 A 1.5 Antrim 2 B 1.0 Antrim 2 C 0.0 Antrim 2 D 0.5 Arenac 1 A 1.0 Chippewa 6 A 3.0 Chippewa 6 B 3.0 Chippewa 7 A 0.0 Chippewa 7 B 0.0 Chippewa 7 C 5.0 Chippewa 7 D 0.5 Crawford 1 A 1.0 Crawford 1 B 1.5 Crawford 1 C 1.0 Crawford 1 D 0.5 Crawford 2 A 8.0 Crawford 2 B 1.0 Crawford 2 C 0.5 Crawford 2 D 2.0 Crawford 3 A 1.0 Crawford 3 B 1.0 Crawford 3 C 1.0 Crawford 3 D 0.0 Crawford 4 A 0.0 Crawford 4 B 0.0 Crawford 4 C 0.0 Grand Traverse 1 A 0.0 Grand Traverse 1 B 0.5 Grand Traverse 1 C 0.5 Grand Traverse 1 D 0.5 Grand Traverse 2 A 1.0 Grand Traverse 2 B 1.5 Grand Traverse 2 C 1.0 Grand Traverse 2 D 1.5 Kalkaska 1 A 1.0 Kalkaska 1 B 0.5 Kalkaska 1 C 0.0 Kalkaska 1 D 0.5 Kalkaska 2 A 1.5 Kalkaska 2 B 0.5 Kalkaska 2 C 1.0 142 Table C3 (cont’d). % 25-Micron 0.0 0.5 0.5 0.5 3.5 2.0 3.5 11.5 1.0 1.0 1.5 3.5 0.5 0.0 1.5 0.5 2.0 1.0 1.5 1.0 3.0 1.0 1.0 1.0 0.5 0.5 1.0 1.5 2.0 3.0 2.0 1.5 1.0 2.5 2.5 2.0 0.5 1.5 1.0 0.5 34.5 41.0 Nd-‘d-‘mmmmaaathNNNA-s-s-s wwwN—s—s—s—s 3 Schoolcraft 1 Tuscola 1 Tuscola 1 A B C D A B C A A B C D A B C D A A B C D A B C D A B C D A A B C D A A B C D A A B 143 Table C3 (cont’d). Location Sample % 25-Micron Clay Wexford 1 A 2.0 Wexford 1 B 0.5 Wexford 1 C 2.0 Wexford 1 D 3.0 Wexford 2 A 0.5 Wexford 2 B 11.0 Wexford 3 A 10.5 Wexford 4 A 9.5 Wexford 4 B 3.5 Wexford 5 A 0.5 Wexford 5 B 0.5 Wexford 5 C 0.5 Wexford 5 D 1.0 Wexford 6 A 0.5 Wexford 6 B 2.0 Wexford 6 C 7.5 Wexford 7 A 0.0 Wexford 7 B 0.0 Wexford 7 C 1.0 Wexford 7 D 0.5 Wexford Sand Co. 1 0.0 Wexford Sand Co. 2 0.5 Wexford Sand Co. 3 0.0 Wexford Sand Co. 4 0.5 144 Appendix D Table D1: Sample location ranking criteria. Higher ranks identify locations that are more suitable to the foundry industry based on these criteria. Cat_agory Criteria Rank All samples subangular 75-99% of samples subangular Grain Shape 50-74% of samples subamgular 25-49% of samples subangular <25% of samples subangular All samples 47-53 At least 2 samples 47-53 Grain Fineness Number (GFN) All samples 44-56 At least 1 sample 47-53 None of the above Mean standard deviation 03 Mean standard deviation 3-6 Standard Deviation of GFN Mean standard deviation 6—9 Mean standard deviation 912 Mean standard deviation >12 Mean acceptable sieves 9 Mean acceptable sieves 6.0-8.9 Sieves Mean acceptable sieves 3.0-5.9 Mean acceptable sieves 0.1-2.9 Mean acceptable sieves 0 Mean clay content 0.0% Mean clay content 0.1-1.0% 25-micron Clay Mean clay content 1.1-2.0% Mean clay content 2.1-3.0% Mean clay content >3.0% OnhO-bOdeAOde-hO—KNW#CAchhO-ANOO# pH pH 6.5-7.8 pH <6.5 or >7.8 Acid Demand Value (ADV) ADV $10.00 ADV >10.00 145 LIST OF REFERENCES Alliance for the Great Lakes, 2007. About the Alliance. http: //www . greatlakes . org/; accessed 24 August 2007. American Foundry Society, 2004. Mold and Core Test Handbook. American Foundry Society, Des Plaines, IL. American Foundrymen’s Association, 1912. Molding Sand Tests. Transactions of the American Foundrymen's Association, 21. American Foundrymen’s Association, 1924. Tentatively adopted methods of tests and resume of activities of Joint Committee on Molding Sand Research. American Foundrymen's Association. American Foundrymen’s Association, 1944. Foundry Sand Testing Handbook. American F oundrymen's Association, Chicago, IL. Anderton, J .B. and Loope, W., 1995. Buried soils in a perched dunefield as indicators of late Holocene lake level change in the Lake Superior basin. Quaternary Research, 44: 190-199. Arbogast, A.F., Scull, P.S., Schaetzl, R.J., Harrison, J ., Jameson, T.P., and Crozier, S., 1997. Concurrent stabilization of some interior dune fields in Michigan. Physical Geography, 18: 63-79. and Jameson, T.P., 1998. Age estimates of inland dunes in East-Central Lower Michigan using soils data. Physical Geography, 19: 485-501. and Loope, W., 1999. Maximum-limiting ages of Lake Michigan coastal dunes: Their correlation with Holocene lake level history. Journal of Great Lakes Research, 25 (2): 372-382. , Wintle, A.G., and Packman, S.C., 2002a. Widespread middle Holocene dune formation in the eastern Upper Peninsula of Michigan and the relationship to climate and outlet-controlled lake level. Geology, 30(1): 55-58. , Hansen, EC, and Van Oort, M.D., 2002b. Recontructing the geomorphic evolution of large coastal dunes along the southeastern shore of Lake Michigan. Geomorphology, 46: 241 -255. , Schaetzl, R.J., Hupy, J .P., and Hansen, EC, 2004. The Holland Paleosol: an informal pedostratigraphic unit in the coastal dunes of southeastern Lake Michigan. Canadian Journal of Earth Sciences, 41(11): 1385-1400. 146 Attfield, R., 2003. Environmental ethics: an overview for the twenty-first century. Polity Press, Cambridge, 232 pp. Ayres, Lewis, Norris, and May, Inc. and Chapman, M.J., 1978. An economic study of coastal sand dune mining in Michigan, Report of Investigation 20. Michigan Geological Survey, Michigan Department of Natural Resources, Lansing, MI. Barnes, B.V. and Wagner, W.H., Jr., 1981. Michigan trees. A guide to the trees of Michigan and the Great Lakes region. University of Michigan Press, Ann Arbor, MI. Blewett, W.L., 1994. Late Wisconsin history of Pictured Rocks National Lakeshore and vicinity. Pictured Rocks Resource Report #94—01 , National Park Service, Washington, DC: US. Government Printing Office and Winters, H.A., 1995. The importance of glaciofluvial features within Michigan’s Port Huron Moraine. Annals of the Association of American Geographers, 85(2): 306-319. and Reick, R.L., 1987. Reinterpretation of a portion of the Munising Moraine in Northern Michigan. Bulletin of the Geological Society of America, 98(2): 169- 175. Brown, G.G., 1936. Molding sands of Michigan and their uses. Michigan Geological Survey, Pub. 41(Geol. Ser. 35). Buckler, W., 1979. Dune type inventory and barrier dune classification study of Michigan‘s Lake Michigan shore. Michigan Geological Survey Report of Investigation 23. Calver, J .L., 1940. Roundness of grains in western Michigan dune sands. Michigan Academy of Sciences, Arts, and Letters, 25: 465-471. Calver, J .L., 1946. The glacial and post-glacial history of the Platte and Crystal Lake depressions, Benzie County, Michigan. Michigan Geological Survey, Occasional Papers on the Geology of Michigan for 1946, Part II. Cowles, H., 1899. The ecological relations of the vegetation on the sand dunes of Lake Michigan. Botanical Gazette, 27: 95-117, 167-202, 281-308, 361-391. Deevey, ES. and Flint, RR, 1957. Postglacial hypsithermal interval. Science, 125 (3240): 182-184. 147 Delcourt, P.A., Petty, W.H., and Delcourt, HR, 1996. Late Holocene formation of Lake Michigan beach ridges correlated with a 70-yr. oscillation on global climate. Quaternary Research, 45: 321-326. Derouin, S.A., Lowell, T.V., and Hajdas, I., 2007. Landscape evolution and deglaciation of the Upper Peninsula, Michigan: An examination of chronology and stratigraphy in kettle lake cores. Journal of Great Lakes Research, 33(4): 875- 886. Devall, B. and Sessions, G., 1985. Deep Ecology: Living as if Nature Mattered. Peregrine Smith Books, Salt Lake City, UT. Dorr, J .A. and Eschman, D.F., 1970. Geology of Michigan. University of Michigan Press, Ann Arbor, MI. Dow, K.W., 193 8. The origin of perched dunes on the Manistee moraine, Michigan. Michigan Academy of Sciences, Arts, and Letters, 23: 427-440. Dow, K.W., 1940. Some examples of ventifacts from Sleeping Bear Point, Leelanau County, Michigan. Michigan Academy of Sciences, Arts, and Letters, 25: 473- 476. F arrand, W.R., 1988. The glacial lakes around Michigan. Michigan Department of Environmental Quality, Geological Survey Division, Lansing, MI. and Bell, D.L., 1982. Quaternary Geology of Michigan (1 :500,000 scale map). University of Michigan Press, Ann Arbor. and Eschman, D.F., 1974. Glaciation of the Southern Peninsula—A Review. Michigan Academician, 7(1): 31-56. Folk, R.L., 1955. Student operator error in determination of roundness, sphericity, and grain size. Journal of Sedimentary Petrology, 25: 297-301. Fraser, G.S., Larsen, CE, and Hester, NC, 1990. Climate control of lake levels in the Lake Michigan and Lake Huron basins. Geological Society of America Special Paper 251. 67-74. Goldsmith, V., 1985. Coastal Dunes. In: RA. Davis, Jr. (editor), Coastal Sedimentary Environments, 2nd edition. Springer-Verlag, New York, pp. 303-3 78. Goodpaster, KB, 1978. On being morally considerable. Journal of Philosophy, 75(6): 308-325. Green, R.M., 1977. Intergenerational distributive justice and environmental responsibility. BioScience, 27(4): 260-265. 148 Harman, J .R. and Arbogast, A.F., 2004. Environmental Ethics and Coastal Dunes in Western Lower Michigan: Developing a Rationale for Ecosystem Preservation. Annals of the Association of American Geographers, 94(1): 23-36. Heinrich, E.W., 1979. Economic geology of the sand and sandstone resources of Michigan, Report of Investigation 21. Geological Survey Division, Michigan Department of Natural Resources. Hofstetter, L.O., 1948. Trends in the foundry sand industry. California Journal of Mines and Geology, 44: 9-36. Johnson, D., 2005. Ford Motor Company. Personal communication. Kelly, R.W., 1971. Geologic sketch of Michigan sand dunes, Pamphlet 5. Geological Survey Division, Michigan Department of Natural Resources, Lansing, MI. Lake Michigan Federation, 1999. Vanishing Lake Michigan Sand Dunes: Threats from Mining. Lake Michigan Federation, Chicago, IL. Larson and Schaetzl, 2001. Origin and evolution of the Great Lakes. Journal of Great Lakes Research, 27(4): 518-546. Lewis, J ., 1975. Michigan's Industrial Sand Resources, Circular 11. Geological Survey Division, Michigan Department of Natural Resources. Loope, W.L. and Arbogast, A.F., 2000. Dominance of a ~ 150 year cycle of sand supply change in late Holocene dune-building along the eastern shore of Lake Michigan. Quaternary Research, 54: 414-422. Marrone, R., 2005. Ford Motor Company. Personal communication. McLaws, I.J., 1971. Uses and Specifications of Silica Sand: Research Council of Alberta Report 71-4. Research Council of Alberta. Michigan Center for Geographic Information, 2005. The Michigan Geographic Data Library. http: //michigan . gov/cgi; accessed 12 March 2005. Michigan Department of Environmental Quality, 1979. Generic Environmental Impact Statement Sand Dune Mining. Michigan Department of Environmental Quality, Geological Survey Division, Lansing, MI. , 1987. Bedrock Geology of Michigan (1 :500,000 scale map). Michigan Department of Environmental Quality, Lansing, MI. 149 , 1989. Atlas of Critical Dune Areas. Michigan Department of Environmental Quality, Lansing, MI. , 2000. Sand mining in Michigan (presentation). Michigan Department of Environmental Quality, Geological Survey Division, Lansing, MI. , 2003. Political Townships Containing Critical Dune Areas. Map. Michigan Department of Environmental Quality, Lansing, MI. 2006. Sand Dune Sand Produced in Michigan: 1978-2005. Michigan Department of Environmental Quality, Lansing, MI. Michigan Department of Natural Resources, 1984. Action Plan Sand Dune Protection and Management Program, Preliminary Report. Michigan Department of Natural Resources, Lansing, MI. , 1985. 2706—Sand Dune Management and Protection—Department Operations. Issued June 14, 1985. Michigan Department of Natural Resources, Lansing, MI. Michigan Environmental Council, 2006. Developing our coastlines: four Michigan communities take stock of their Great Lakes assets. Michigan Environmental Council, Lansing, MI. Michigan Legislature, 1970. Shorelands Management and Protection Act, Act No. 245, Public Acts of 1970. State of Michigan. , 1972. Soil Erosion and Sedimentation Control Act, Act No. 347, Public Acts of 1972. State of Michigan. , 1976. Sand Dune Protection and Management Act, Act No. 222, Public Acts of 1976. State of Michigan. , 1994. Natural Resources and Environmental Protection Act, Act No. 451, Public Acts of 1994. State of Michigan. , 2001. Michigan’s Natural Resources and Environment: A Citizen’s Guide. Legislative Service Bureau, Science and Technology Division, Michigan Legislature, Lansing, MI. Michigan Natural Resources Commission, 1985. 2706 - Sand Dune Management and Protection - Department Operations. Michigan Natural Resources Commission, Lansing, MI. Moldenke, R., 1930. The Principles of Iron Founding. McGraw-Hill, New York, 654 pp. 150 Naess, A., 1973. The shallow and the deep, long-range ecology movement: a summary. Inquiry, 16: 95-100. National Climatic Data Center, 2003. United States Climate Normals. National Oceanic and Atmospheric Administration. http://lwf.ncdc.noaa.gov/0a/climate/normals/usnormals.html; accessed 18 August 2006. Natural Resources Conservation Service, 2006. Web Soil Survey. United States Department of Agriculture. http: //websoi lsurvey . nrcs . usda . gov/app/; accessed 20 August 2006. Olson, J ., 1958a. Lake Michigan dune development 1: Wind velocity profiles. Journal of Geology, 66: 254-263. , 1958b. Lake Michigan dune development, 2: Plants as agents and tools in geomorphology. Journal of Geology, 66: 473-483. , 1958c. Lake Michigan dune development, 3: lake level, beach, and dune oscillations. Journal of Geology, 66: 345-351. Palmer, C., 1997. Environmental Ethics. ABC-CLIO, Santa Barbara, CA, 192 pp. Parkes, W.B., 1950. Composition and Properties of Sands (Resources Report 274). Journal of Resources and Development, British Cast Iron Resources Association. Pebbles, V. and Thorp, S., 2001. Waste to Resource: Beneficial Use of Great Lakes Dredged Material. Great Lakes Commission, Ann Arbor, MI. Rieck, R.L. and Winters, HA. 1993. Drift volume in the southern peninsula of Michigan—a prodigious Pleistocene endowment. Physical Geography, 14(5): 478-493. Ries, H. and Rosen, J .A., 1908. Foundry Sands. Geological Survey of Michigan Annual Report, pp. 37-78. Rolston, H., 111, 1975. Is there an ecological ethic? Ethics, 85: 93-109. Routley, R., 1973. Is there a need for a new, an environmental, ethic? Proceedings of the World Congress of Philosophy, Varna, Bulgaria, pp. 205-210. Russow, L., 1981. Why do species matter? Environmental Ethics, 3(2): 101-112. Schaetzl, R. and Isard, S., 1991 . The distribution of Spodosol soils in Southern Michigan: A climatic interpretation. Annals of the Association of American Geographers, 81(3): 425-442. 151 and , 1996. Regional-scale relationships between climate and strength of podzolization in the Great Lakes region, North America. Catena, 28: 47-69. and Weisenbom, B., 2004. The Grayling Fingers region of Michigan: soils, sedimentology, stratigraphy, and geomorphic development. Geomorphology, 61: 25 1-2 74. Schlesinger, ME. and Ramankutty, N., 1994. An oscillation in the global climate system of period 65-70 years. Nature, 367: 723-726. Schoeneberger, P.J., Wysocki, P.J., Benham, EC, and Broderson, W.D. (editors), 2002. Field book for describing and sampling soils, Version 2.0. Natural Resources Conservation Service, National Soil Survey Center, Lincoln, NE. Scott, 1., 1942. The dunes of Lake Michigan and correlated problems. 44th Annual Report of the Michigan Academy of Science, Arts, and Letters, 44: 53-61. and Dow, K., 1937. Dunes of the Herring Lake Embayment, Michigan. Michigan Academy of Sciences, Arts, and Letters, 22: 437-450. Searcy, K.B., Wilson, BP., and Fownes, J .H., 2003. Influence of bedrock and aspect on soils and plant distributions in the Holyoke Range, Massachusetts. Journal of the Torrey Botanical Society, 130(3): 158-169. Singer, P., 1975. Animal Liberation: a new ethics for our treatment of animals. Avon Books, New York. Stevenson, EB, 1931. The dunes of the Manistique area. Michigan Academy of Sciences, Arts, and Letters, 14: 475-485. Sundeen, S.P., 1978a. Geologic study of sand deposits in the state of Michigan: Phase 1 final report, Open File Report - OF R 78 03. Michigan Geological Survey, Michigan Department of Natural Resources, Lansing, MI. ., 1978b. Geologic study of sand deposits in the state of Michigan: Phase 2 final report, Open File Report - OFR 78 O4. Geological Survey Division, Michigan Department of Natural Resources, Lansing, MI. Tague, GO, 1946. The post-glacial geology of the Grand Marais embayment. Michigan Geological Survey, Occasional Papers on the Geology of Michigan for 1946, Part 1. Taylor, P.W., 1981. The Ethics of Respect for Nature. Environmental Ethics, 3: 197-218. 152 Thompson, T.A., 1992. Beach ridge development and lake level variation in Lake Michigan. Sedimentary Geology, 80: 305-318. . and Baedke, S., 1995. Beach ridge development in Lake Michigan: shoreline behavior in response to quasi-periodic lake level events. Marine Geology, 129: 163-174. . and Baedke, S., 1999. Strandplain evidence for reconstructing late Holocene lake level in the Lake Michigan basin. In: C.E. Sellinger and RH. Quinn (Editors), Proceedings of the Great Lakes Paleo-Levels Workshop: The Last 4000 Years, pp. 30-34. US. Army Corps of Engineers, 2007. 800 Locks—Sault Ste- Marie. http://www.lre.usace.army.mil/_kd/go.cfm?destination=page&pge_id= 1 3 2 4; accessed 15 August 2007. Van Dijk, D. and Vink, D.R., 2005. Visiting a Great Lakes Sand Dune: The example of Mt. Pisgah in Holland, MI. The Great Lakes Geographer, 12(2): 45-63. Van Oort, M., Arbogast, A. F ., Hansen, E. C., and Hansen, B., 2001. Geomorphological history of massive parabolic dunes, Van Buren State Park, Van Buren County, Michigan, Michigan Academician, 33:175-188. Venteris, ER, 2007. Three-dimensional modeling of glacial sediments using public water-well records: An interpretation of interpretive and geostatistical approaches. Geosphere, 3(6): 456-468. Wyckoff, M., 1986. Managing Sand Dune Development in Michigan: State and Local Options. Michigan Department of Natural Resources, Lansing, MI. 153