ABSTRACT AN ANALSIS OF PLEISTOCENE SEDIMENTS IN AN AQUIFER RECHARGE AREA, KALMAZOO, MICHIGAN by Patricia A. Asiala Travis The recharge area on Station 9 of the City of Kalmazoo's aquifer system has been the subject of intensive study by the city and the state and federal geologic surveys. Numerous test wells have yielded data on flow, recharge, relative permeability, and gross com- position of the sediments in the aquifer. A detailed study of the individual sediments involved was undertaken by the author. The objective was to establish a positive or negative relationship between the physical and mineralogical parameters of the sediments and their permeability. The samples were Ro-tapped to determine size dis- tribution, checked for roundness and sphericity, and a mineralogical analysis of the clay fraction was made for samples having a high clay percentage and high permea- bilities. Standard deviation, sphericity, roundness, medians, skewness, and kurtosis measures were computed for all samples. In addition, x—ray analysis was done for eleven samples. Patricia A. Asiala Travis Most of the sediments involved displayed the typical size distribution of well sorted channel sands. A few tills were present as well as a few clay hardpan sediments. The median size was found to be the key factor in evaluating the permeability of most of the samples. The bimodal distributions proved to be difficult to evaluate by median size alone and the percentage of clay and sorting values had to be taken in consideration. Roundness and sphericity varied only slightly and no correlation between these factors and permeability could be found on the scale of this study. Relationships might be found if a larger area were covered. Clay-size percentages of over 10% were able to reduce the effective permeability to such a low level that all other factors were unable to overcome its effects. The mineralogical composition of the clay fraction indicates that these samples have at least two different origins and possibly different depositional histories. Two samples were predominantly kaolinite which displayed a high degree of crystallinity. These samples had little or no other clay minerals in this size interval. Most of the other samples had a heterogeneous mixture of kaolinite, illite, montmorillonite, and other clay minerals liberally mixed with non-clay minerals such as quartz. Patricia A. Asiala Travis The clay mineral suite in glacial sediments might prove to be a very effective tool in determining provenance or mode of deposition when more data is available for comparison. The presence or lack of clay minerals may be significant in itself. AN ANALYSIS OF PLEISTOCENE SEDIMENTS IN AN AQUIFER RECHARGE AREA, KALAMAZOO, MICHIGAN By Patricia Ann Asiala Travis A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Geology 1966 ACKNOWLEDGMENTS The author wishes to express appreciation for time, materials, information, and pertinent advice as given by Morris Deutsch and K. E. VanLier of the Lansing, Michigan Branck of the U.S.G.S. Groundwater Division, and those members of her committee: Dr. B. T. Sandefur, chairman Dr. H. B. Stonehouse Dr. M. M. Mortland Dr. M. M. Miller Dr. W. J. Hinze The generous permission to use the x—ray equipment and laboratory facilities of the Soil Science Department as granted by Dr. Mortland is sincerely appreciated. ii TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES LIST OF FIGURES LIST OF APPENDICES Section INTRODUCTION Location of Research Area. . Summary of Previous Investigations. Statement of the Problem . GEOLOGIC HISTORY Nature of Bedrock in Research Area. Pre- Pleistocene Erosion . . . Pleistocene Glaciation and Erosion. COMPOSITION AND SOURCE OF GLACIAL SEDIMENTS THE Glacial Tills: Definition and Mode of Deposition. Factors Controlling Composition of Glacial Sediments . . . . . . . . CLAY MINERALS. Clay Mineral Structures The Tetrahedral Sheet The Octahedral Sheet Kaolinite . . Illite Montmorillonite Chlorite Ion of Exchange Absorption of Water. Ion Substitution and Potassium Fixation iii Page ii vi vii H ocasq -q UJFHA Section Page WEATHERING AND ALTERATION OF GLACIAL SEDIMENTS . 23 Mineral Composition of Glacial Sediments . . 23 Environmental Influences on Mineral Composition. . . . . . . . . 23 The Leucocratic Silicates . . . . . . . 24 The Ferromagnesian Silicates . . . . . . 25 The Carbonates. . . . . . . . . . 25 The Layered Silicates: The Clays and Micas . 26 Clay Mineral Formation by Weathering . . . 27 Clay Formation by Recombination . . . . . 28 The Role of Iron . . . . . . . . . . 29 The pH Factor . . . . . 3O Concentration of Ions Due to Leaching . . . 3O PERMEABILITY. . . . . . . . . . . . . 32 Factors Controlling Permeability. . . . . 32 SIZE ANALYSIS . . . . . . . . . . . . 38 Purpose . . . . . . . . . . . . . 38 Method . . . . . . . . . . . . . 38 SPHERICITY . . . . . . . . . . . . . 42 Definition . . . . . . . 42 Factors Controlling Spheric ity . . . . . A2 ROUNDNESS. . . . . . . . . . . . . . AS Definition . - . . . . . . . . . . A5 Rate of Rounding . . . . . . 46 Factors Affecting Roundness Values . . . . A6 STATISTICAL ANALYSIS . . . . . . . . . . A9 Purpose . . . . . . . . . . A9 Methods and Parameters Used . . . . . . 49 THE X—RAY ANALYSIS OF THE FRACTION OF SELECTED SAMPLES . . . . . . . . . . . . . . 55 Procedure . . . . . . . . . . . . 55 Technical Data. . . . . 58 Identifying Characteristics of Selected Minerals. . . . . . . . . . . . 58 IV Section Page RESULTS AND CONCLUSIONS . . . . . . . . . 63 The Importance of the Clay Fraction. . . . 63 THE SHAPE OF THE PERCENTAGE DISTRIBUTION CURVE . 79 PERCENTILE SIZE DISTRIBUTION GRAPHS . . . . . 81 THE EFFECT OF SORTING ON PERMEABILITY . . . . 114 The Effect of Sphericity on Permeability . . 118 The Effect of Roundness on Permeability . . 118 The Effect of MdQ on Permeability . . . . 123 SUMMARY . . . . . . . . . . . . . . 128 SUGGESTIONS FOR FURTHER RESEARCH . . . . . . 131 APPENDICES . . . . . . . . . . . . . . 133 BIBLIOGRAPHY . . . . . . . . . . . . . 136 Table LIST OF TABLES Formulae for Statistical Quantities Used Changes Produced by Heating Some Common Minerals in Clay Fractions. X-ray Diffraction Data Sorting versus Permeability Sphericity versus Permeability Roundness versus Permeability. Permeability versus 9 Median vi Page 54 59 60 115 119 121 125 Figure 10. 11. 12. 13. 1A. 15. l6. l7. l8. 19. LIST OF FIGURES Index Map Showing Location of Kalamazoo Area, Michigan Map of Pumping Station 9 and Sample Sites Block Diagram and Section through Recharge Area . . . . . . . . . Surface Geology of the Kalamazoo Area Preglacial Topography of the Kalamazoo Area Structures of Some Common Clay Minerals. The Effect of Consecutive Particle Sizes on Passage of Fluids . . . . . Flow Sheet of Sediment Analysis Flow Sheet for X-ray Procedure. The Composition of the Clay-Size Fraction as it Relates to Permeability in X-rayed Samples. . . . . X—ray Diffraction Diagram: Sample 5. X—ray Diffraction Diagram: Sample 6. X—ray Diffraction Diagram: Sample l6 X—ray Diffraction Diagram: Sample 22 X-ray Diffraction Diagram: Sample 3A X—ray Diffraction Diagram: Sample 35 X-ray Diffraction Diagram: Sample 37 X-ray Diffraction Diagram: Sample 40 X—ray Diffraction Diagram: Sample A3 vii Page 33 39 56 67 68 69 7O 71 72 73 7A 75 76 Figure Page 20. X—ray Diffraction Diagram: Sample A6 . . 77 21. X-ray Diffraction Diagram: Sample A8 . . 78 22. Percentile Size Distribution: Samples A5, 33; 35, and A6 . . . . . 82 23. Percentile Size Distribution: Samples 27 and A8 8A 2A. Percentile Size Distribution: Samples 20, 21, 7, and 28 86 25. Percentile Size Distribution: Samples 18 and 23 88 26. Percentile Size Distribution: Samples 10 and A2 90 27. Percentile Size Distribution: Samples 37, 38; 8, and 6. 92 28. Percentile Size Distribution: Samples 9 and A3 9“ 29. Percentile Size Distribution: Samples 22 and 25 96 30. Percentile Size Distribution: Samples 1A and AA 98 3l. Percentile Size Distribution: Samples 26 and 30 100 32. Percentile Size Distribution: Samples 13 and 3A 102 33. Percentile Size Distribution: Samples 3l and Al lOA 3A. Percentile Size Distribution: Samples 5, l9; l6, and 29 106 35. Percentile Size Distribution: Samples 15 and I7 108 viii Figure Page 36. Percentile Size Distribution: Samples ll, 2A; 12 and A0 . . . . . 110 37. Percentile Size Distribution: Samples 32, 39; 36 and A7 . . . . . 112 38. Q Sorting Compared to 0 Md . . . . . . 117 ix Appendix A. B. LIST OF APPENDICES Sphericity and Roundness Data Statistical Data for all Samples Used Page 13A 135 INTRODUCTION Location of Research Area The area under study is located l/A mile south of the city limits of Kalamazoo, Michigan, on the west fork of Portage Creek. It is known as Station 9 of the city's aquifer system. The well sites are in Portage Township, Sect. A, T.3S., R.llw., as shown in Figure 1. Summary of Previous Investigations The City of Kalamazoo, the Geological Survey Division of the Michigan Department of Conservation, and the United State Geological Survey Ground Water Branch, Lansing, Michigan, have conducted a co-operative study of the ground-water conditions in the Kalamazoo area since l9A5. A comprehensive report covering the results of these studies has been published as MGS Progress Report #23 (Deutsch, Fan Lier, and Giroux, 1960). Faced with an inadequate supply of water from naturally recharged aquifers, the City of Kalamazoo began digging recharge ponds and developing an arti— ficial recharge system to overcome the deficiency. The program has been notably successful. It has supplemented the water supply to such an extent that in Spite of 1 I’D 0‘! !(0 Is. I "“ll. ’5‘) R- |_?-W AtLaaAN i ~'\ 3.5 .3 § ITEXA§S‘ Ei._~___’_._§7 ESPRA|¥?E II RONDE gs " :1: é¥aosefifi— '" Figure 1.--Index map showing Location of Kalamazoo area, Michigan. '%§RL&§§I CO r. POR — Raaaa *—-——- Rallraada a Lalm N svraama RJIW. 2. “came": ...R-'°Y’.-. lop TEGEIPAMI LIONI \ -—: cALHbuu ..—- KALAIAIOO COUNTY EXPLANATION _..._ County linu __- Township llaaa E35553 City of Kalamazoo MGS Progress Report 23 expanding industrial and domestic usage, no shortage has been experienced. All samples for this thesis come from wells on Station 9. Deutsch, Vanlier, and Giroux (1960, p. 3) describe the earlier studies which have been run on these wells as follows: . . . . source, occurrence, availability, chemical quality, and use of groundwater, included a series of aquifer tests, and also a program of test drilling, financed by the City of Kalamazoo, to determine the thickness and extent of the fresh-water aquifers underlying the city and contiguous areas. Using the information about the geology and hydrology of the area obtained during the study, and new techniques of well development, maintenance, and induced recharge, the city's Utilities Department has since expanded its well and pumping facilities and has developed one Of the largest groundwater supply systems in Michigan. Statement of the Problem Three wells were chosen from those drilled at Station 9 (see Figure 2, p.14). These wells represent a cross-section of a diversion of Portage Creek. They were drilled for the purpose of measuring the amount and rate of recharge to the aquifer under controlled surface conditions. The USGS had sampled these wells for the purpose of obtaining permeability data on the various types of glacial sediment involved. Samples were taken with a well bailer. The well casing was driven past the zone to be sampled and the material within the casing was brought to the surface with a bailer. The permeability STATION 9 ' f 0 wells sampled 0 other wells 0 OO ’\ ”7m. ‘ 74L: Portage Creek Figure 2.--Map of Pumping Station 9 and Sample Sites. AMER. WATER WORKS ASS'N P. 186 U1 tests were run by the USGS Hydrologic Laboratory in Denver, Colorado. Using the samples thus obtained and the permeability data available, the author proceeded to determine the sedimentary parameters and clay mineralogy of the samples for the purpose of determining any relation— ship between these factors and the measured permeability. unit I I--II~d-dI—h -’ock diagram and section through recharge area. GEOLOGIC HISTORY The Michigan Basin has existed as a structural low since possibly Precambrian time. Evidence of deposition prior to late Cambrian has not been uncovered to date. An interval from the Cambrian to the Pennsylvanian was marked by deposition of shallow marine sediments accompanied by periodic subsidence. The result was a sequence of marine beds similar to a number of nested saucers. The sediments range from normal marine limestones and shales to thick evaporite sequences and some coal beds. The time interval from the early part of the Pennsylvanian to the early Pleistocene is represented by few sediments. Those which have been sampled show a mixed marine- terrestrial character. Recent studies on the fossil spore and pollen content of these formations tentatively place the formations in the Late Jurassic (Geologic Section of Michigan 196A, Michigan Department of Conservation, Geological Survey). Nature of Bedrock in Research Area The Mississippian Goldwater shale directly underlies the glacial deposits in the Kalamazoo area. It is a bluish shale with low permeability, contributing little 7 I .H. . .. ENF’E .. . . £91,! . \Ilv .- n.- adtd - an? Ida-Ia a: Cu 3: Alana, $414.3}— HNPEI v. 9‘50: ~ flit... v buzz- J/fihfi -. y in": nuw mow IX PLANA‘I’ION I Ma any and alayay llll at mum In aarmaablllty. Laaally and!» Ianaaa al sauna-Ila and naval whaan aaualn ”IMO“. Arr-«nu and an on"! lama lam-um m lav-a. um. and clay al valallvaly In ”mummy. ml aaaaam Pamaaala «M and araval law- I“ "an mammal ”a an UJJaalaalaal Swan lamlhlaal «mm u "an! Lavavan. FIGURE A.--Surface geology of the Kalamazoo area. ”A . 1 ' O 1: III’ I L a 2’ l 8. 0'.’ l I I 1 . I ' I I \ a) I I ‘ \w E I l .A~dL—'ro"\ I : EXPLANATION V550"" N Contour on lha badraak aorta“, daahad whara Infarrad Contour lolarval 80 “at Datum la moon aaa laval 0.00-00. Traoa of major burlad PIalatoaaoa vallay d--- ‘l’raoa of molar llama pra- Plalataaaoa vallay , O ! Illaa - . Nola! Coalaara ara largoly Natalia and an aabiocl to . MGS Progress Report 2 3 chaooa as additional m bacama availabla. Figure 5.--Preglacial tOpography of the Kalamazoo area. 9-H pun-0 10 water to the present aquifer system. The small amount it does contribute is high in CaSOu and generally unpotable. Pre—Pleistocene Erosion Following the Jurassic and preceding the Pleistocene epoch, a long interval of erosion is postulated. There is no evidence for deposition of beds during this interval and this indicates either erosion Evidence for erosion may be noted of the underlying bedrock surface. by large river systems whose main followed the present axial trends Huron, and Erie. Many subsidiary and/or non-deposition. in the dissected profile This surface is marked channels appear to have of Lakes Michigan, channels also cut the bedrock. Since erosion rarely takes place so that its products are completely removed one might infer that a residue of depositional material awaited the arrival of the first glacier. Pleistocene Glaciation and Erosion During the Pleistocene the Michigan Basin and its environs were subjected to repeated glaciation. The environs were subjected to repeated glaciation. The Nebraskan, Kansan, and Illinoian glacial intervals preceded the Wisconsinan and may have encompassed much of the area. During the Wisconsinan four major glacial advances and retreats apparently removed all recognizable remnants of previously deposited unconsolidated material, 11 as well as some of the bedrock. These sediments were either removed from the area or reworked to such an extent that they can not now be distinguished. The area sampled is within the boundaries of the Wisconsinan glaciation and is covered with deposits of glacial material ranging from poorly sorted tills to stream channel deposits which are very well—sorted (Figure 3, p. 6; Figure A, p. 8). Three principal ice lobes, the Lake Michigan Lobe, the Lake Erie Lobe, and the Saginaw Lobe affected the Southern Peninsula of Michigan. The Lake Michigan and Saginaw Lobes traversed the area in question. Anderson (1955) delineated the approximate outline of these lobes and other lobes in the Central Lowland of the United States. He also presented a discussion of the general pebble lithology of the drifts involved. COMPOSITION AND SOURCE OF GLACIAL SEDIMENTS Glacial Tills: Definition and Mode of Deposition According to R. F. Flint (1957 p. 108) till is a very heterogeneous sediment which may be compared to a rapidly deposited alluvium such as a flood deposit. Till is typically unsorted or poorly sorted. Some tills are predominantly boulders, some have clay as the dominant constituent: others have only sand, silt, and pebbles, with neither clay or boulders. In Michigan the first two types are common, the latter rare. It is often difficult to distinguish the precise mechanism which laid down a particular sediment. This is particularly true in glacial materials. Melt—water is always presert at the edge of a glacier as material is deposited. This water has a tendency to remove fine material or at least stratify it, whereas large fragments mixed with fines are commonly left where the ice melts as shown in the area under study. Factors Controlling Composition of Glacial Sediments The proportion of a particular rock or mineral in any of the size fractions is a function of the following: 12 13 The distance from source. The areal outcrop of the source rock. The resistance of the parent rock to erosion. The durability of the constituent minerals and rock fragments during transport. (Flint, 1957, pp. 127-129) 1‘:me An outcrop of a particular rock situated close to the depositional area will contribute a larger per- centage of its material to the final sediment than an outcrop of a similar rock situated several miles farther from the site. A source rock outcrop having twice the area of a similar source rock will contribute a proportionately larger amount to a deposit than a smaller outcrop, other factors being equal. However, the factors of resistance to erosion, resistance to mechanical abrasion during transport, and the distance from source to deposit, so far outweigh the size of the original outcrop as to make it only of minor significance. For example, a relatively small outcrop of a quartzite may contribute boulders to a deposit several tens of miles away whereas a less resistant shale may not be carried more than a few hundred feet. It follows that a rock which is very dense and resistant to abrasion will contribute less material than a similarly sized outcrop of a soft, easily eroded rock at the outset. However, the more resistant rocks may travel long distances without losing their individual characteristics and less resistant rocks are rapidly 1A broken down. R. F. Flint (1957, pp. 123-126) cites several instances of rocks traveling several hundred kilometers from their source. THE CLAY MINERALS* Clay Mineral Structures All crystalline clay minerals belong with the micas in the structural class of the phyllosilicates. All clay minerals have a layered structure consisting of two or more sheet-like arrangements of atoms. There are two basic types of sheets; the tetrahedral, with a general composition (SiOu)_u and, the octahedral, with a compo- sition of either (Mg3(OH)6) or (Al2(OH)6). These basic units are arranged in varying number and order, with interlayer cations, to form the different clay minerals. The Tetrahedrel Sheet In the tetrahedral sheet, three oxygen atoms form a "base" for the silica tetrahedron and are shared in three directions with adjacent silicon atoms, forming a sheet. The fourth oxygen is not shared within the sheet and occurs directly above the silicon atom in the tetrahedron. The Octahedral Sheet Six (011)"l ions in the octahedral sheet assume a 6-fold coordination about a central cation, either (Al+3) 2>. or (Mg+ Hydroxyl ions are shared at the corners to *Grim, 1953. 15 16 form a continuous network in two dimensions. Since the hydroxyls on the corners of the octahedra are shared with adjacent octahedra, the ratio of central cation to the octahedron is either 2:1 or 3:1. In the case of the trivalent Al+3 ion the ratio is 2:1 and the corresponding sheet is known as a dioctahedral sheet. In the case of 2 ion (also Fe+2) the ratio is 3 Mg+2 ions per the Mg+ octahedron and the sheet is referred to as a trioctahedral sheet. The dioctahedral sheets are also referred to as gibbsite sheets since their composition (Al2(OH)6) is identical with that of the mineral gibbsite. Similarly the trioctahedral sheets are often referred to as brucite sheets since their composition is the same as that of brucite. The tetrahedral and octahedral sheets combine in varying sequence and number by replacement of the (OH) ion in the latter by the non—planar oxygen in the former. Kaolinite Kaolinite consists of a tetrahedral silica sheet and an octahedral gibbsite sheet and possesses no inter- layer cations. This mineral has an exposed face of hydroxyls which are subject to ion exchange by replace- ment of some of the hydrogens by other cations. 17 111159 Illite has a unit structure consisting of a gibbsite or brucite sheet intercalated between 2 silica tetrahedral sheets. These units are stacked and joined by having K+1 1 ions tend to ions in the interlayer positions. These K+ "fix" the dimensions of the structure in the c—axial direction and limit its expandability. Illite has less ion exchange capacity than montmorillonite. Some illites closely approach the structure of muscovite as the substi— tution of Al+3 for Si?“ in the silica sheet increases and the cations between sheets are bound in a 12-fold co- ordination (Dana, 1959, p. ASA). Montmorillonite Montmorillonite has a layered unit structure similar to that of illite, i.e , two silica sheets and an inter— calated gibbsite or brucite sheet. The unit structure is joined to another similar unit by interlayer cations, Ca+2, Mg+2, Na+l, etc , which are subject to ion exchange. These cations are held between the surfaces of two primary units by charges arising from substitution in the tetra— hedral sheet and the octahedral sheet. The silicon in the tetrahedral sheet may be partially replaced by aluminum giving rise to excess negative charge. Magnesium or ferrous iron may substitute for part of the aluminum in the octahedral layer giving rise to excess negative l8 charge. This excess charge is somewhat weakened by the distance from the surface at which it is neutralized, so cations are usually not tightly held and structural units tend to be rather loosely bound together. Chlorite Chlorite consists of two triple-layered units with an interlayer brucite sheet. It is non-expandable. It has a relatively low cation adsorption capacity. Iron may substitute for part of the Mg+2 in the brucite sheet varying the chemical composition of the mineral. Ion Exchange Positive ions are often adsorbed on or between the structures of clay minerals. In the upper profiles and in soil horizons these ions are often organic as well as the usual alkaline earth and alkali metals. The adsorbed ions may be exchanged with other ions in the environment of the particles. The ease of exchange varies with the mineral involved and the concentration of exchangeable ions in the environment. Sites of Exchange There are two principal exchange sites: 1. Those cations held between units may be exchanged with others which satisfy the same charge imbalance and whose dimensions allow them to fit into spaces between surfaces of units. l9 2. On the edges of the particle are a number of broken bonds where sheet—like layers have been disrupted. Ions may be adsorbed and exchanged at these sites under favorable conditions. Adsorption of Water Montmorillonite, vermiculite, and some less common clay minerals may adsorb water molecules in interlayer positions. There is always some water adsorbed on the broken bond surfaces of the particle, but it is relatively thin compared to that of the intercalated layer (Grim, 1953, p. 162). The water layer which is adsorbed directly on the clay surface varies in thickness from 3 to 10 angstroms. Most clay mineral researchers agree that the adsorbed water is more dense and viscuous than that of ordinary water. The precise reason for this difference and its exact nature have not been established. The thickness of the adsorbed layer is known to vary with the presence of adsorbed cations and to a different degree with different cations. According to Grim (1953, pp. 176-177) the effect of various cations is dependent on, (1) the ability of a cation to stabilize the distance between layers of the mineral, (2) the hydration of adsorbed cations, and (3) the size of the ion. Multi— valent cations tend to stabilize structure more readily than most monovalent cations with the exception of potassium. Hydration of adsorbed cations produces an oriented group of water molecules around each cation 20 and influences the thickness of the adsorbed water layer. Larger ions tend to disrupt the molecular net of water and probably hinder its formation, conversely smaller ions would aid in the formation of a water net. Organic complexes and other materials are often adsorbed at the exchange sites. Ion Substitution and Potassium Fixation In micas, substitution of up to half of the Si“1 by Al+3 may take place in the tetrahedral sheet. This produces a strong charge imbalance which is compensated for by the positive charge of potassium or other alkali ions held tightly between units. In clay minerals most of the substitution takes place in the octahedral sheet. This also produces an excess negative charge, but it acts through a greater distance than a charge arising from substitution in the tetrahedral sheet. The charge arising in the tetrahedral sheet has less effect and can not hold cations as strongly. Therefore, most clay minerals tend to lose their interlayer cations rather easily. Potassium, and in some cases, magnesium, tends to form a more stable mineral than either sodium or calcium. Potassium has an ionic radius which conforms closely to the dimensions of holes in the surfaces of the silica sheets. Once adsorbed, it is removed with difficulty 21 and has the net effect of removing the exchange sites which it occupies from participation in the ion exchange process. Adsorption of potassium also stabilizes the intercunit dimension of the structure and forming a pseudo-mica structure commonly called illite. The non—replaceable adsorption of potassium ions is known as potassium fixation. 22 Figure 6.--Structures of some common clay minerals. [W mm... 7.21 0.0.99.0.09 “ushommaolo (Pettilohn. 1957) . )()< >1>(' » , p. 132. L b O I A1AISiu010)(OH)8 (Damal, p.' A61)‘ l clad; lanaia-a- A ILLITE . _ O KyAJ-L‘(S.'I.8._yAly)020(OH))4 (OH) K Al -F -M - . 10.01 . 14 VI 5' 65' 314 Me;6)(818_y Aly)020 (Pettijohn, 1957) p. 13A . C, OH . 31 4 ° A1, Mg, Fe O o O‘ Na, Ca,.H2O, etc. K, H2O TM MONTMORILLONITE 9.3.21.4 , ‘ .9 anyfluoua IKE; "?5\ . . 0 Mason, 1952 Figures 1 and 18 p. 13A ' WEATHERING AND ALTERATION OF GLACIAL SEDIMENTS Mineral Composition of Glacial Sediments Since glacial sediments in Michigan have a multiple origin and heterogeneous mineralogy, their alteration must be considered as a matter of individual components as well as a whole. Four mineral groups make up the bulk of glacial sediments. These are the leucocratic silicates, the ferromagnesian silicates, the carbonates, and the layered silicates. These minerals may be present as mono- mineralic grains or as aggregate rock fragments. If they occur as the fragments the structure of the parent rock will affect the disaggregation and the disintegration of the component grains. The original texture and structure affects all mineral grains involved, since grain-size is controlled by the crystal size in the original rock. Structures such as jointing, schistosity, and slaty cleavage affect the rate of weathering. Environmental Influences on Mineral Composition Minerals and rock particles are weathered as a function of their environment and their resistance to it. The physical environment and their resistance to 23 2A it. The physical environment may encompass a number of variables including temperature, pressure, and movement of water. The chemical environment involves the nature of the fluids with which the particles come in contact, the Eh, and the pH. Interactions between the weathered products of minerals are common and can directly influence the rate or type of weathering. The Leucocratic Silicates The feldspars and quartz comprise the bulk of the minerals in outwash materials and sandy tills in Michigan. These minerals are present in all size fractions. Under the microscope quartz is seen as clear, clean, angular to subangular fragments. In part it may be derived from pre-existing sediments such as the Marshall sandstone. The rounded grains exhibit overgrowths indicating the sedimentary origin of part of the quartz. Other particles are sharply angular, with no rounding or overgrowths and are assumed to be from igneous or metamorphic rocks. Quartz in the Pleistocene sediments has probably under- gone little weathering. It is possible minute quantities have been dissolved but there is no evidence to substan- tiate this. Any colloidal silica present in the clay fraction is probably derived from the decomposition of other silicates. 25 Plagioclase and potassic feldspars are present. With the aid of a microscope these minerals are seen as angular to rounded grains so heavily clouded by alteration that characteristics normally used in their identification (twinning, birefrigence) are badly obscured. The Ferromagnesian Silicates This discussion will exclude the biotites since they are included in the layered silicates. The other common ferromagnesian silicates belong to two principal groups, the amphiboles and the pyroxenes. The less common ferromagnesian silicates affect the bulk composition to no appreciable degree. The amphiboles and pyroxenes are present as fine slightly altered particles. They are accessory components in the sediments with which this research deals. The_Carbonates The carbonates, dolomite and calcite, are the other major mineral group to be discussed. These minerals are usually altered by leaching by rain and ground-water. The Ca+2 and Mg+2 ions are removed in solution or adsorbed at the surfaces of the layer silicates. The carbonate ion combines with the hydrogen ion from the water and carbon dioxide from the air to form a soluble bicarbonate, susceptible to leaching by circulating waters. 26 In general calcium carbonates are more soluble than magnesium carbonates and are more readily removed from the leached profile. The Layered Silicates: the Clays and Micas Clay minerals and micas are another major group of minerals found in glacial sediments. Their alteration may be considered from two viewpoints; as degradational or aggradational. The micas alter to clays by degradation, (hydration and removal of cations). Clay minerals may dissociate to colloidal and ionic units or may gain cations which stabilize and restore order to their structures. Clays may also form from ions and colloids in solution. Clays resulting from these varied processes are often impossible to differentiate from each other. In the sediments with which this paper deals two suites of clay minerals are probably present. Some of the clays were formed by the alteration of mineral particles since and during transport, and from the weathering profiles which develOped on the parent rock prior to glaciation. The other suite belongs to a prior cycle of sedimentation. This suite includes the clay minerals of the marine sediments which underlie the glacial tills and which were not exposed to surface weathering conditions between the time of their burial and their removal by glacial scouring. The clay minerals in the bedrock would be expected to be 27 close to equilibrium prior to removal and should exhibit more structural perfection and larger particle size. The formation of a clay mineral is simply the assumption of the crystal form having energy most compatible with the environment. The structure of clays is, there- fore, rather heterogeneous because of the variety-of environmental conditions where clay is found. Substitu- tional and transitional structures are common and the task of classifying clays is rather complex. Clay Mineral Formation by Weathering During weathering, minerals such as the micas lose interstitial cations, gain water, and assume a clay mineral structure and cell dimensions. The weathering of three dimensional structures is similar but involves an inter— mediate step of ionic solution, followed by a constructive crystallization to the clay most nearly approximating the composition of the solution from which it is crystallizing under the prevailing temperature and pressure controls. According to Grim (1953, p. 3A3) a basic igneous rock will weather to montmorillonite if the weathering profile is not leached to a high degree and Mg+2 is retained in the profile. Kaolinite will form if the environment has a high degree of leaching. Grim also suggests that kaolinite will not form in an environment having high + Ca 2 content (pp. 351). This would Open the question of 28 whether kaolinite found in calcareous tills is primary or secondary. It would also explain the high content of kaolinite in the weathered profile above the unleached calcareous zone. Clay Formation by Recombination Jenny (1951) gives the alteration of silicates as a combination of two processes, hydration and hydrolysis. He attributes the formation of secondary clay minerals to the recombination of sheets of polyhedra broken loose by these two processes. Correns (19A9) indicates that the formation of secondary silicates is primarily a precipi- tation of ions or recombination of ions in a more stable configuration, rather than an incomplete breakdown of silicates and recombination of polyhedral fragments. Correns found that aluminum and silicon were, for the most part, in true ionic solution before the formation of clays. Fredrickson (1951, p. 122) visualizes the decomposition of silicates as a process of ion substitution, whereby the water absorbed on the surface of an albite particle, for instance, will assume a crystalline orientation and exchange H+1 ions for Na+1 ions in the particle. This is accomplished by the H+1 ion being incorporated into the albite lattice and upsetting the neutrality of the lattice. The crystal will try to become neutral and to lose the 29 excess charge by expelling Na+2 ions. The net result is an expansion of the crystal, an increase in chemical reactivity, and a collapse of the lattice. This produces gels of different compositions depending on the Al:Si ratios of the parent material. Whichever of these processes is operative during weathering, the end products are colloidal alumino- silicates, gels, iron oxides, and cations which are either removed in solution or incorporated into a new silicate. The Role of Iron The cations involved in weathering are those making up the majority of the constituent minerals in the earth's +2 +1 +1 +2 8‘ ’ 1+8, Al+3, Fe+3, Fe+2, Ca ', N Mg , etc. crust; S , K Of these, the Fe+2 and Fe+3 ions are unusual in that they are not easily incorporated into new silicate lattices once they have been weathered out of the parent mineral. These ions usually occur as the precipitated hydrated iron oxides, limonite, hematite, and goethite. This is partly due to the very slight solubility of iron at the pH of the normal weathering environment. Ferric hydroxide pre— cipitates at a pH of about three and ferrous hydroxide at a pH of 5.1 (Rankama and Sahama 1950, p. 66A). The pre- cipitation of iron is also aided by the presence of calcareous material found in many glacial tills. 30 The pH Factor The pH of a weathering profile is affected by (l) the oxidation of sulfides, (2) the presence of organic acids, or (3) the release of metallic ions from the break— down of minerals in the sediment. Glacial sediments contain only small amounts of sulfides so that source of H+ ions is negligible. Humic acids are also usually confined to soil profiles and do not affect the majority of the sediment. However, organic matter possesses the capacity for ion exchange, whereby H+ ions may be exchanged for other ions in the environment, thus increasing the H+ ion concentration of the interstitial fluids and lowering the pH. The release of metallic ions decreases the pH only if the ions are removed by leaching. The amount of water moving through a particular system is therefore of utmost importance in the alteration of a material. It controls the amount of organic acids which move downward in the profile and the removal or non-removal of ions in solution, and thereby the equilibrium of the system. The amount of water supplied to the system is a function of the per- meability of the material. It follows that permeability is then a factor in the weathering of a glacial profile. Concentration of Ions Due to Leaching As the individual cations (Na+l, K+l, Ca+2, Mg+2, etc.) are moved out of the system or participate in the adsorption 31 phenomenon of ion exchange, other less soluble or less reactive cations are concentrated with respect to the total composition. Iron and silicon move out only with difficulty and at a slow rate. Al+3, Fe+3, Fe+2, and Ti+2 tend to form insoluble compounds and are not affected by leaching under ordinary circumstances in a normal pH range. PERMEABILITY Factors Controlling Permeability Three principal quantities control permeability in any unconsolidated material. They are temperature, hydraulic gradient, and the coefficient of permeability of the material. Water temperature is inversely pro- portional to viscosity and in the aquifer under study the temperature does not vary sufficiently to seriously affect the flow of water. The hydraulic gradient is assumed to be a constant at the time of testing although it varies with the amount water being supplied at different times of the year. The coefficient of per- meability is dependent on a number of subsidiary factors which exert their influence in differing degrees to control the flow of water through the system. These factors are particle size, sorting, shape, roundness, grain surface, amount and type of clay, stratification, and porosity. Krumbein and Sloss (1956, p. 91) state that permeability is strongly influenced by particle size. Their argument is based on the occurrence of larger openings between particles of large sizes as compared to more numerous but smaller openings between particles of small sizes. Since fluids would have less friction 32 33 with large particles and molecular attraction between particles and liquid would be less due to decreased surface area, permeability should be greater. This is easily observed in a collection of uniform spheres where modifying influences are at a minimum. Figure 7 illustrates the case in natural sediments. D i “ 0A. . “a .r FIGURE 7.--The effect of consecutive particle sizes on passage of fluids. 3A In Figure A particles of two consecutive size intervals are present and large particles cannot maintain grain to grain contact, leaving wide inter—particle channelways. Figure B has particles of two sizes which are non- consecutive. Here the larger particles may be in contact and many of the channelways are constricted so that even small amounts of clay or silt will clog the passage. There is more frictional drag on the fluid trying to pass through these small interstices. The effect of two conse- cutive sizes is noticed most in the medium to fine sand ranges since the inter-particle passages are relatively small. When one deals with very coarse sand and gravels the effect is not important because the passages are large to begin with. Sorting affects permeability in proportion to the average grain diameter. Fraser found in a study of porosity and permeability of unconsolidated material, that with an increase in grain size, the conditions slowly approach those existing in a one component system. Fraser (1935, p. 959). In general, poorly sorted sediments are less permeable than those which are well sorted. Porosity is affected by particle shape and surface texture. Krumbein and Sloss (1956, p. 90) state that porosity is affected by shape as well as the uniformity of size and shape. Porosity is found to be greater in loosely packed sediments, poorly sorted sediments and 35 in finer grained sediments. In the latter the degree of random orientation and packing were partially responsible for differences in total porosity. Pettijohn (19A9, p. 89) gives the following relation— ship for porosity and permeability which includes the factor of specific surface: K = 2 x 107 x O x l__. (l — 0) S2 K - Permeability in darcys O — Porosity 2 3 S - Specific surface (cm. /cm. ) Specific surface is calculated from the size analysis, assuming all grains are spheres. Pettijohn states that this relationship holds true for unconsolidated natural sands. The type and amount of clay present in a sediment will affect the permeability of the sediment. Fine clay particles are strongly attracted by other mineral particles and tend to coat these larger particles. Clay particles also have an attraction for water molecules which causes some of the interstitial fluid of the sediment to be more tightly held. They reduce the size of the interstitices of sediments by filling them in, either partly or completely. Clays may be in a dispersed state due to the ion content of the interstitial water. This would cause clogging of the interparticle passages and reduce permeability. If the clay is in a flocculated state, 36 the permeability may be little affected since the clay will not move with the interstitial fluid. The property of ion exchange in clay minerals has a profound influence on the permeability of a sediment. W. P. Kelley (1955) found that calcium saturated clays tended to be granular and relatively porous, while Na+2 saturated clays are dispersed and relatively impermeable. Grim (1955) states that the degree of hydration and the cation-particle spacing determines the amount of influence the particular cation exercises on a particular sediment. The degree of hydration of Na+1 is greater than that of Ca+2 or Mg+2. They in turn are more hydrated than Fe+3 or Al+3. The degree of hydration seems to determine the ease of replacement and the strength of the bond between cation and clay particle. Na+1 tends to be very loosely held and easily replaced, whereas H+1, Fe+3, and Al+3 are firmly held and difficult to replace. The hydration of Na+1 is sufficient to break apart clay mineral layers, producing a dispersed, finely particulate clay, which clogs interstices and coats particles, reducing permeability. Ca+2 and Mg+2 do not have this effect on the crystal structure and are associated with larger particles and less dispersion. Ca+2 and Mg+2 are more tightly held on adsorption surfaces and are not as readily available in solution as hydroxides to act as 37 diSpersants. Na+1 which is more loosely held, forms a dilute hydroxide which acts to diSperse fine particles. Small amounts of Na+2 in a system may substantially reduce the permeability. Other ions are not nearly as effective in this respect. Highly substituted clays will also be more likely to break up and release ions into solution. The Substitution of ions in either the octahedral or tetrahedral layers weakens the bond between the sheets and clears the way for mechanical and chemical breakdown of the lattice. (Johns and Jonas, pp. 163-171). The fine particles resulting will be subject to increased ion exchange due to increased surface. They are also more likely to go into solution and supply additional ions. The particles remaining will be dispersed and clog interstitial openings, affecting the permeability. The type of clay mineral present affects per- meability since clay minerals which participate in ion exchange to the greatest degree are also most sensitive to changes in the composition of the ground-water and to the action of weathering solutions. SIZE ANALYSIS Purpose The analysis of grain size in a sediment is based on the premise that grain size distribution is not only characteristic of a particular type of sediment but a function of the depositional and transportational environ- ments. Grain size, like sphericity and roundness, is influenced by hardness, mineral composition, source, degree of weathering, and other such factors. The parameters of size analysis are usually statistical in nature and measure the absolute percentage in specific size grades as well as cumulative percentages and distri— bution of sizes, range of sizes, etc. Thus size analysis is the primary step in a complete statistical analysis. It supplies the raw data for statistical inferences regarding the nature of the sediment. Method Size analysis is commonly done by sieving or settling. A scale of grade sizes is adopted and suitable screen sizes or time intervals chosen to fit a particular mathematical relationship. Several grade scales have been devised, riotable among them the Wentworth and Atterberg scales. The choice of grade scale is not important in itself as long 38 39 WEIGH COMPLETE SAMPLE RO—TAP COMPLETE SAMPLE WEIGH SIZE INTERVALS CHOOSE SAMPLES FOR CLAY ANALYSIS SEPARATE CLAY FROM SILT BY SETTLING AND DECANTATION I' MOUNT CLAY ON CERAMIC PLATE I I I X-RAY Figure 8.-—Flow sheet MOUNT SANDS ON SLIDES ROUNDNESS STUDIES SPHERICITY I STUDIES I ROUGH MINEF [ALOGICAL ANALYSIS of sediment analysis. A0 as intervals are chosen which are Spaced to give an adequate representation of the size distribution of the sediment. The result of the analysis is plotted as a cumulative curve and enough points must be identified to give the curve validity. Size of sand grains is a measure of the average grain size between two limits, the aperture of the screen through which it has passed and the retaining screen. In settling a time interval marks the limits. In a natural sediment the particles are not perfect spheres and their deviation from the spherical influences their passage through a screen. Flat or long particles will find it difficult to pass through a screen as their most stable position is parallel to their greatest dimension. Each grade size is there- fore a composite of several shapes of an average size corresponding to the intra-screen interval. One must speak, therefore, of a size interval rather than a single size. Sieve analysis is used for sizes larger than silt. Below silt size, efficiency and accuracy decrease to the point where this method is inapplicable. Settling techniques have been used for sand through clay sizes but it is usually more efficient to sieve sands mechanically. Settling may be done by a number of methods, including decantation, rising current elutriation, pipette analysis, etc. The author chose decantation to separate the silt A1 and clay fractions. This followed weighing, sieving, and other procedures (see Figure 8). Sphericity and roundness studies were done under a microscope using comparison charts rather than the more lengthy projection method. Accuracy is not sufficiently different to warrant the time spent on the longer method. The mineralogical analysis was also done under the micro- sc0pe to check the possibility of mineralogical variations which might correlate with the permeability behavior. None were found and the mineralogy was suprisingly similar. The small areal extent of the sampling and the fact that the samples are from different horizons representing different depositional intervals precludes the usefulness of heavy mineral studies. SPHERICITY Definition Wadell (1932) gives the formula for sphericity as follows: Sphericity = Volume of the particle f Volume of a circumscribing sphere The ratio of surface area to the volume (sphericity) of a particle is a clue to the depositional and transportational conditions through which the particle has passed. It helps control the response of a particle to lifting forces. Sphericity also influences the selective transport of fragments to a large degree. When considered with size and density it should reflect changes in the environment in which the particle was deposited. Factors Controlling Sphericity Original sphericity is probably controlled by struc- tural weakness and grain size in the parent rock. Mineral composition is also important since pronounced cleavage, distinctive fracture, or pronounced twinning may influence the original sphericity. The Sphericity of a particle at any given time after release from the parent rock is a function of some of its A2 A3 inherent nature and its environment. Original sphericity, time, grain size, genetic type, the method of size reduction, and lineations or other grain weakness, are all partly responsible for the sphericity of a particle. Timp.--Time may influence sphericity in one of two ways: either accentuating it or decreasing it. A mineral with a pronounced cleavage would tend to be less spherical over a period of time due to continued breakage along planes of weakness. Actinolite would be an example. In contrast a mineral with equal or nearly equal bonding strength in all directions would tend to be rounded off and assume a more spherical appearance with time. Hardness, brittleness, and cleavage all influence the ultimate sphericity. Grain Size.-—Grain size will affect sphericity, for very small grains will have less mass and will tend to be cushioned by any available fluid, therefore will not have the impact nor the surface for impact to be effective on collision. They also will have less structural weaknesses. In general small particles have less sphericity than larger ones for these reasons. One of the most obvious illustrations of this is the pronounced angularity of silt particles, especially those in loess. Genetic Type.--Genetic type is particularly important to quartz sphericity. The particle shape of quartz varies greatly when found in a gneiss, a granite, or a sandstone. Size will vary along with the amount and type of lineations, AA presence of overgrowths, and the average angularity. The average angularity. The strained quartz grains from gneisses and schists will have more inherent weaknesses than those from an unmetamorphosed parent rock. Sedimentary overgrowths are common on quartz grains in sandstones, increasing the sphericity of the particle. Method of Size Reduction.—-Size reduction is most effective on larger particles and therefore the method to which a particle is subjected will produce the most change in the Sphericity of larger particles. The method may be chemical solution, abrasion by other particles, or impact breakage. Chemical solution is the only method which is noticeably effective on small particles because of their large surface area. Small particles lack the necessary weight to be effectively reduced by impact and do not offer enough surface or enough resistance to abrading particles to be effectively reduced by this means. Inherent Weaknesses.--Lineations such as bubble trains and strain axes are possible sites for breakage in a grain. Particles with a high percentage of lineations will break and maintain a low sphericity. ROUNDNESS Definition Wadell (1932) defines roundness as the average radius of the corners and edges of a particle divided by the radius of the maximum inscribed circle. Average radius of corners and edges Roundness = Radius of maximum inscribed circle The maximum roundness is assigned a value of 1.0 for a perfect sphere or such forms as cylinders terminated by hemispheric ends. Values less than 1.0 indicate a particle which is less than a perfect sphere or has corners or edges which do not have one spherical cross-section. Krumbein and Sloss (1956) found that roundness is a function of the distance a particle travels as well as the rigor of the transportational medium. Roundness does not directly influence the settling velocity of particles and therefore transportational and depositional environments are not shown by particle rounding unless they affect the rate of wear of particles, (Krumbein, 19Al, pp. 6A-72). Pettijohn states (19A9, p. 53): "The roundness of a clastic particle sums up its abrasional history." A5 A6 Rate of Rounding Sharp corners and edges wear off more rapidly than rounded irregularities of the particle surface and a particle will undergo a rapid stage of rounding until these sharper projections are worn down. The rate of rounding will decrease considerably as the fragment becomes smoother. Factors Affecting Roundness Values Selective transport does not cause roundness but tends to emphasize the rounded nature of a particular sediment. For instance, a sediment may have a charac— teristic roundness which is not due to its abrasional history. Selective transport of rounded material on a beach may cause concentration of rounder particles nearer shore where another agent, such as wind may concentrate them further by size selection. The final sediment may be a sand dune or beach sand with a high degree of round- ness which does not indicate its true abrasional history (Anderson, 1926). The absolute size of a particle influences the degree of rounding. Particles of large size are more likely to be rounded than particles in the fine sand sizes. Particles may lose much of their angularity by solution and chemical attack in the very fine sizes (below 0.59 mm.). In general this is a minor factor in the rounding of mineral grains (Marshall, 1929). Larger particles are also more likely A7 to have structural weaknesses than smaller particles. If such a large particle breaks, the fragments are usually angular and may be subjected to an increase in angularity due to crushing between larger particles. LeRoy (1950, p. 200) suggests using fine sand sizes for roundness determinations because the rounding takes place more slowly in this size interval and a more accurate indication of the abrasional history of the sediment may be obtained. The mineralogical composition must also be considered. Quartz is often used as the criterion for roundness studies. Other minerals such as the carbonates may be more'common in a particular sediment but since they are susceptible to rounding by both abrasional and chemical means, they might give a misleading representation of the amount of abrasional history the sediment has sustained. Quartz is one of the best standards since it has the requisite hardness, ubiquity, and resistance to solution to serve as a good indicator. The rounding of quartz under normal sedimentary conditions is rather slow and the presence of well rounded quartz grains usually indicates more than one cycle of erosion. Quartz derived from previously existing sediments often exhibits overgrowths which increase the apparent roundness but which must be ignored if one desires the true round— ness of the particle. A8 Kuenen (1956) gives the order of rounding suscep— tibility as follows: very low . . . flint, radiolarite, chert, agate low . . . quartz, rocks, rhyolite medium . . . granite, gabbro, diabase, graywacke, some limestone very high . . . other limestones, lava, obsidian Rounding is a function of the following factors: (1) Hard- ness of the mineral or rock; (2) Brittleness; (3) Original jointing, etc., strains, or cleavages; (A) Grain size; (5) Time since release from parent rock; and (6) Rigor of transportational and depositional media. STATISTICAL ANALYSIS Purpose Statistical measures were applied for the specific purpose of checking any relationships between permeability and statistical data on the samples. W. C. Krumbein and P. D. Trask pioneered the application of statistical measures in geology. The application of statistics has been expanded considerably by numerous authors and an ex- tensive list of papers on the uses, limitations, and appli- cations of statistical measures to problems in sedimentation and other phases of geology has been compiled. Problems having large amounts of data which are cumbersome to treat by conventional analysis have been expedited with the use of statistical analysis. Statistics must be applied with the same discretion which one uses when applying other methods of study, in that all types of problems do not lend themselves to successful statistical study. There are many statistical parameters which as yet have found no practical use in the natural sciences. Methods and Parameters Used The following measures were chosen by the author for use in this study: 1. the 5th, 16th, 50th, 8Ath, and 95th percentiles, A9 50 2. the Q arithmetic mean, 3. the 9 median, A. the 9 standard deviation (sorting), 5. the 0 skewness measure, 6. the 2nd 9 skewness measure, and 7. the Q kurtosis measure. The Percentiles The measures are those used by Inman (1952). Inman chose the 16th and 8Ath percentiles because they are one standard deviation unit on either side of the mean, whereas the 5th and 9th percentiles are quite close to two standard deviation units on either side of the mean. Inman states that there is little appreciable error in measurement of percentiles one standard deviation away from the mean. He also chose the 5th and 95th percentiles rather than the 2 1/2 and 97 1/2 percentiles (two standard deviations) because of the ease of measurement of the former and an appreciable reduction in error. The Phi (0) Scale.--According to Krumbein and Pettijohn (1938, p. 233) the 9 scale is a scale where each interval is the equivalent of one Wentworth grade and of equal length with every other unit. The scale increases with decrease in size of the particle diameter. It has the advantage of direct plotting on arithmetic graph paper and points may easily be read to tenths and hundredths. 51 Statistical measures may be applied to such graphs and the answers are in Wentworth units. The 0 Arithmetic Mean (M9) The arithmetic mean is the average grain size in a distribution (see Table l on page 5A). The 0 Median (MdQ) The phi median is the 50th percentile. It is less affected by skewness than the arithmetic mean and is more useful when emphasis is on the most abundant size, according to Inman (1952). Inman also contends that the arithmetic mean is of greater use where a number of means from sedimentary layers must be averaged. In a symmetrical curve the phi median and the arithmetic mean are equal. Inman defines the phi median as: ". . . the diameter value which divides the frequency curve into two equal areas, and the phi arithmetic mean as the diameter of the center of gravity of the frequency distribution." The 0 Standard Deviation (SQ) Sorting or the measure of dispersion is one-half the distance between the 16th and 8Ath percentile diameters. This is a measure of the deviation of the curve from the norm, in Wentworth units. ‘ r ... p draw 0* (l) g). u.- L‘). I In I'\\ u.“ A.” V‘ (I) .’\' 1" (I) (I) (I) ’( 1 ' 'J ‘- (1' “r Ca 52 The 0 Skewness Measures (“01 and ag2) Skewness is theoretically the difference between the mean and the median, but to keep it independent of the deviation of a distribution, it is divided by the standard deviation. Values of skewness are negative if the dis- tribution is skewed to larger sizes and positive if the distribution is skewed to fine sizes. Two skewness measures are employed; the first reflecting the skewness of the bulk of the distribution, and the second, the skewness within the tails of the distribution. The 0 Kurtosis Measure (8g) Kurtosis is another measure often calculated for sediments. Inman defines it as: "the ratio of the average spread in the tails of a distribution to the standard deviation of the distribution." The application and signi- ficance, if any, of this measure has yet to be demonstrated. It is a measure of the peakedness of a curve. Values have been assigned for kurtosis; 0.65 being normal, values less than 0.65 indicating that the tails have less spread than normal and the curve is more peaked than normal. Conversely, values greater than 0.65 indicated a flatter curve with more spread in the tails than normal. 53 Conversion Table for Wentworth Units to Phi Units* Wentworth Q 256 mm. -8 128 -7 6A -6 32 -5 16 —A 8 -3 A -2 2 -1 1 O 1/2 1 l/A 2 1/8 3 1/16 A 1/32 5 l/6A 6 1/128 7 1/256 8 1/512 9 1/102A 10 *Krumbein, 1950, p. 567. According to Krumbein (1950) the class limits of the Wentworth scale form a geometric series, in which each succeeding grade is one-half as large as its predecessor. This series may be written as: -U D = 2 D = Diameter in millimeters fl = An exponent such that any value of D has a corresponding sign in terms of 0 A negative sign is used so that any size less than unity has a positive sign in 9 terms. 5A TABLE l.--Formulas for statistical quantities used. The Arithmetic Mean (M0) Mg = 1/2 ($16 + 88A) The Phi Standard Deviation (SQ) 89 = 1/2 (”an ' ”16) The Phi Skewness Measures 1/2 (“16 + flan) ‘ Mda Mm ‘ Mdz lst Moment afll S = SQ 0 ) - Md G = 1/2 (95 + 995 2nd Moment ag2 S Q The Phi Kurtosis Measure = 1/2 (995 - 95) —.S U B 9 SO THE X—RAY ANALYSIS OF THE CLAY FRACTION OF SELECTED SAMPLES Procedure Samples which exhibited both a permeability between 100-1000 meinzers and a clay-silt percentage between two and nine per cent were selected for analysis. This range encompasses all of the permeability anomalies observed. The fine fractions were separated from the sands and gravels by the use of Ro-Tap shaker. The clays and silts were further separated into sizes of greater than and less than 20 microns by settling. This was done by dispersion of the sediment in de-ionized water and putting several drops of the .lN NaOH in the suspension until the fine material remained in suSpension. The larger particles were allowed to settle for one minute and then the liquid suspension was decanted. This separated the remainder of the fine material. The decanted liquid was then reshaken and allowed to settle for five minutes. The liquid was again decanted and the residue was resuspended. After another five minute settling period the material still in suspension was decanted and combined with that which had been removed before. The residue was resuspended, filtered, dried, and weighed. This fraction was comprised of 55 56 l. Ro—Tap complete sample Clays and Silts l Disperse (de—ionized water, r’ We gh Settle (l min.) Decant suspension Liquid I Sands and Gravels I .lN NaOH) Residue of Fine Sand I E ' . Rj-suspend Sittle (5 min.) Decant I Suspension Residue \ Re-suspend r I Glycerate (few drops) Settle (2A hours Deposit on ceramic plate Leach (3 increments .lN Mg 01) Settle (5 min.) Depant I Liquid ReJidue Addlwater (de-ionized) R se (Several increments distilled water, 3% glycerol by volume) Filfer Dry (Room temperature, first in air Dry then in dessicator 2A hours) I Weigh X l -|ray SILT > 20u Liach (3 increments .lN KCl) Rinse (As before) H4at (2 hours, 110°C. electric furnace) X-ray L_ Optional Step \ V r (2 hours, 550°C.) 4 \ Heat (350°C. 2 hours) X-iay Figure 9.--Flow Sheet of x—ray procedure. 57 particles of silt size, i.e., larger than 20 microns. The remainder of the material, still in suspension, was treated with a few drops of glycerol and allowed to stand over— night. It was then deposited on a ceramic plate by using a vacuum pump, and a plate holder with a well. This produced a sample in which the basal planes of these essentially platy minerals were closely aligned with the horizontal surface of the plate. The oriented sample was leached with three increments of .1N MgCl (method according to M. M. Mortland, Dept. Soil Science, Michigan State Univ.). This was followed by rinsing with several incre— ments of distilled water, 3% glycerol by volume. The sample was then dried at room temperature, first in air, then in a dessicator for 2A hours and x-rayed as a Mg+2 saturated, glycerol solvated, oriented aggregate. After the initial x-rays the cation was varried, using the same method as with the Mg+2 but by leaching with .lN KCl. The sample was heated for 2 hours at 110° C. in an electric furnace. This was followed by x-rays as a K+ saturated aggregate. A third x-ray was run after the sample had been heated for 2 yours at 550° C. Kaolinite decomposes after 550° C. to a form known as metakaolinite which has no strong x-ray reflections. This facilitates its differ— entiation from chlorite and vermiculite. In a few cases 58 an x-ray diffractometer graph was run after heating to 350° C. to differentiate between chlorite and vermiculite. Technical Data The x—ray diffractometer graphs were indexed using Jackson (1956), Brown (1961), and The Index for the X-Ray Powder Data File. All x-ray work was done on a Norelco x-ray diffraction unit, using a Geiger counter goniometer and recorder. Cu Ka radiation was used with a Ni target. The diffraction unit was Operated at 15 ma. and 35 kv. The recorder was operated at 1 degree per minute and a run from 2 degrees = 26 to about HO degrees was taken for each sample. This gave d spacings from about 20 angstroms to 2.2 angstroms. Identifying Characteristics of Selected Minerals A number of characteristic changes occur on heating of minerals which commonly occur in the clay fraction. These may be identified and used to differentiate the various minerals which are usually too small to be identified by other means. In the case of minerals belonging to the layered clays a loss of interlayer water is a common result of heating above 550° C. This causes a reduction of the c dimension of the unit cell which is reflected in the x-ray pattern. Other hydrous minerals will either lose their water of hydration in an incremental 59 pattern or in a gradual fashion, each of which is charac- teristic for the particular mineral involved. The consti- tuent minerals of the clay fraction and their changes upon heating and treatment of different types may be seen in Table 2. TABLE 2.--Changes in structure and composition accompanying heating of some Common minerals in clay fractions. Mineral Change Kaolinite Decomposes after heating to 550° C. to meta-kaolinite, a semi—amorphous form with little reflection. Chlorite Stable but has an intensification of its 14 A line after heat treatment to 550° C. Vermiculite Some vermiculites change from IA 3 to a 10 or 9.6 A Spacing with heating above 300 degrees. Montmorillonite Expands to 18 A with glycerol treat- ment and collapses to 10 A when heated to 550 degrees. Some varieties collapse at 300 degrees. Chert Some crystallization at 550° C. Cristobalite Should be some inversion to quartz 'at 575° C. Dolomite Partial to complete decomposition at 350° C. leaving a broad band of low intensity. Gibbsite Largely destroyed at 300° C. Goethite Largely destroyed at 300 degrees, but begins to decompose at about 250° C. One hour at 300° C. will decompose it to hematite. 60 mm mO.: om :O.H no Hm.m om mm.m OOH Om.: OOH mH.: mpHHmpouwHAOIm om mm.H NH mm.H om mm.H mm mm.: OOH m:.m Nphmzalm OOH :m.m Nphmsalu MH OO.H o: ow.H MH mm.H O: OH.N OOH mw.m OOH mm.m mpHEOHoQ mH OH.m mH mm.m OOH :O.m mpHOHmo ACOHpmHUMHux mmv OO Hm.m OO mm.m om mm.: om 00.: OOH m.:H OOH 3.:H mpHHSOHELm> OOH m:.H OO mm.m OOH+ mH.b UmLmUhomHU OOH wm.m OOH wm.m memln OOH mH.> mpHCHHomm Ob mm.m OOH NO.m OO m:.H om w:.: OOH om.: OOH 53.: OOH m.mH OOH Om.mH OOH OO.mH muHCOHHHLOEpCOE om 5:.m om OO.~ om om.m om :m.H OOH om.: mHOHcoocHHo OOH m.MH muHLOHSO III mm.m III mm.m OOH mm.m OOH mm.m III no.3 III mm.: OOH OO.: OOH O:.: III o.OH III H.OH OOH m.m OOH o.OH mpHHHH H C\c H c\© H C\o H C\c HmhmcHz .mpme consumammfln gag-x--.m mamas 61 Hm Hm.q Hm OO.m OOH Om.H eOmOHO OH Hm.m OH OH.O OOH OO.m mcHHmOOmz OO MH.m OOH NH.H OOH Om.O OOHEHOHHH Om mH.H O: OO.m OOH HH.O Omm . mommm OO OO.m OH OO.m OOH mO.m m m m O OO OO.m OOH OH.: O m: . O O O OH OOH OH.O HmOOOOm OOHOOOOO OO OO.H OO Hm.m OOH OO.m OHHOOEOm mm mm.m Om Om.: OOH mm.m meHHooHOHz HIHmOmOHmm OO ON.H OO mm.m OOH Om.m OOHOHO om mm.H Hm mm.m Om mm.m OH HO.H OOH mm.m OOH Om.m OHHHOOOHOOH mouwz HmdeHmm H c\O H cxc H C\O H c\O HmpmcHz OmscHOqOOLL.m mHmae .mcHH pmmmcopum on» HOH OOH mo mSHm> m so Ummmn mcHH map Ho HpHmcmch one u H mlmwmlm u C\© Ho O ch Om u c on wcHUHooom mEoppmwcm CH mammmH moprmH Ho wcHoQO O one u U .COHpmHUMH Omso Ho magma CH mum OmHmHomdm mmHzmmzpo mmoch mwcHUmmH HH¢ ”meoz 62 OH mm.: Om Om.m OOH mm.m mpHumcwmz mo m:.m Om HH.m OOH mO.H mpHHHm Om Hm.m mm OO.: mm m0.0 OOH mm.m OOH mm.m OOH O0.0 mpH>oomzz OOH OH.m OOH m~.m OOH ::.H OOH :H.m OOH NH.m OOH OO.w muzmHDChom mm :m.m OOH mN.H OOH am.m mpHC®EHH H c\O H axe H :\O H c\O HmnmcHz " 1" OOOOHOOOOnu.m HHO m s U) Ln E m m H H Ln C) 0 ° N O O O ' ° 08 I ° Ln 0 ' ° ° \0 (\J I I o 0 m :T a) [—1 m U\ U\ I I H I I I I I m m mx I I I I I I I \O \O 01 LG I C) O O O C) o o r—I (\1 Ln . . . . - o o o o o H m _‘3' CD \O H Sample 27: Sample u8: lOOO meinzers 840 meinzers Permeability .715 1.02 Sorting 1.00 .68 9 Median Figure 23.——Percentile size distribution: Samples 27 and 48. 86 Figure 24.——Percentile size distribution: Samples 20, 21, 7, and 28. A. Samples 20 and 2l.--With a pair of unimodal distributions the median size is the indi- cator of permeability. The larger median of sample 20 combined with very good sorting produces a better permeability value than the smaller median and better sorting of sample 21. B. Samples 7 and 28 --Both of these samples have a fairly large median, good sorting, and good permeability. Note that sample 7 has an open- ended curve. It also has better sorting in its middle range than sample 28. Although sample 28 has a considerable larger median, the distribution of material into 2—3 classes about the mean causes more interstitial blocking and a lowering of permeability. 87 Sample 20: Sample 21: 630 meinzers MOO meinzers Permeability -75 .65 Sorting 60 r 1.12 1.73 G Median 50 MO 30 2O 10 L E' 0%“] s \0 LO O O (\J Ln 0 ' . H (\l O o o ' N V o 9 Ln 0 o o o \O m p I I . . (\I :r 00 H I H I I I H I I I l I H (\I Ln I I I I I l O (I) \O (\l LG I O O O O ' \ O '—I (\I Ln 0 o o °- \0 >5 0 . o ,—I m :1- 00 H m H c) 50 _ 40 3O 2O 10 L 0% Figure 2A.—-Percenti1e size distribution: Samples 20, 21; 7 and 28. 88 Figure 25.-—Percenti1e size distribution: Samples 18 and 23. A. Samples 18 and 23.--The similarity of these two curves is evident but sample 18 has a median in the coarse sand size range as I contrasted to the median of sample 23 in I the fine sand size range. The sorting values are nearly identical and there is no signifi— cant difference in clay percentage. The permeability is a direct function of the median diameter. 60 - 50 P uo— 89 - I 1 , 0%(1) Ln E <1) 01 E H H LO Q ' (\J O O O ° 025 I ' Ln 0 ° ° 0 \O I I ’ (\I :T 00 H In In I I H I I I I m m In I I I I I I \O \D (\J Ln I O O O O O O H (\I Ln 0 o o o o o I o H (\J :- a) Sample 18: Sample 23: 1300 meinzers 140 meinzers Permeability 1.515 1.48 Sorting .18 2.12 0 Median Figure 25.—-Percenti1e size distribution: Samples 18 and 23. 90 Figure 26.--Percentile size distribution: Samples 10 and 42. Samples 10 and 42.--These two samples illus— trate a case where the median size is not a good indicator of the probable permeability. Sample 10 follows the usual pattern of a unimodal distribution. It has a median in the medium sand size range, good sorting, and a very good permeability. Sample 42 has an almost identical median but there the resemblance ends. The curve is an open-end distribution and the median is not a good indicator of permeability. The shape of the curve shows a high clay—silt fraction which is probably the major factor in reducing permeability to the low level of 24 meinzers. The heterogeneous mixture of sizes allows much opportunity for blocking of interstitial passages. —_——-_——— 91 80 - 70 ' 60 — 50 n 40 ~ 30 “ 20 H \ 10'*\ 0% a I. 5' m m E H r—I L0 0 0 (\J O O O ' 08 I ° L(\ O ° ° ° \0 l I 0 - (\l 2' CO H U\ uw I I H I I I I m m In I I I I I I KO '\O (\J Ln I C O O O O O H (\I LI'\ 0 o . o . o . . . I—I (\J :1” (13 Sample 10: Sample 42: 640 meinzers 24 meinzers Permeability .225 1.985 Sorting 1.64 1.68 0 Median Figure 26.-—Percentile size distribution: Samples 10 and 42. 92 Figure 27.——Percentile size distribution: Samples 37, 38, 8, and 6. Samples 37 and 38.—-Both of these samples have a low permeability which is primarily due to the high percentage of clay and silt which they have and to the small size of most of their material. The sorting values are relatively poor but the bulk of the material is in the small sand sizes and this concentration produces the slight amount of permeability present. Samples 6 and 8.--The smaller median of sample 6 is a reflection of the clay—silt fraction present. The bulk of the material in both samples may be seen to be in the same size interval. The sorting of sample 6 is much better than that of sample 8. Sample 8 is open-ended toward the larger sizes and samply 6 toward the smaller sizes. This accounts for the higher permeability of Sample 8. 93 Sample 37: Sample 38: 130 meinzers 65 meinzers no _ 1.69 2.79 1.85 2.00 30 _ z'“~ / \ Permeability Sorting 0 Median / 10 ./ /’ l I I‘~—_J—— I ° 0% m 0:: (\J LIN O . H (\I O O O ' (\J . , Ln O o o o \o m LG I I ‘ (\J :1' (I) H I (\I I I I —I I l I I I \o m U\ I I I I I I o C) \O (\J LG I O O O O ' . C) .r—I (\I Ln 0 ° 0 - \0 v . - . I—I (\I :r (I) I—1 /"\ __.. ”’ \f’ I I I I I Sample 8: Sample 6: 350 meinzers 260 meinzers Permeability 1.71 .51 Sorting .94 1.89 9 Median Figure 27.—-Percentile size distribution: 38; 8 and 6. Samples 37, 9U Figure 28.-—Percentile size distribution: Samples 9 and 43. A. Samples 9 and 43.-—The sorting index of these samples shows a vast difference. The medians are relatively close. Sample 9 is a unimodal curve with excel- lent sorting, little clay or silt and little coarse material. Sample 43 has 4—5% clay and silt, much gravel, moderately poor sorting, and a considerably better permeability. The shape of the distribution curve indicates the reason for the differ— ence. While the clay-silt and fine sand percentages give sample 43 a relatively small median size, the mean size for sample 9 lies in the medium sand size range and the mean for sample 43 is in the coarse sand range. This difference is sufficient to give sample 43 the edge. 95 0% I’i l I I\-—-I-——I-——-I— . In E (\I H U'\ C) ’ N C) O C) ° N I ° Ln O o o . \O \O I I ' 0 (\I :1” (I) H o In I I H I I I I . m U\ I I I I I I V \O (\I LG I O O O O O H (\I m ' ' O o - - - - H (\I z 00 Sample 9: Sample 43: 630 meinzers 820 meinzers Permeability .245 1.81 Sorting 1.51 1.32 G Median 1.63 .58 0 Mean Figure 28.——Percentile size distribution: Samples 9 and 43. 96 Figure 29.——Percentile size distribution: Samples 22 and 25. A. Samples 22 and 25.-—These samples illustrate the effect of clay and silt on the perme- ability of well-sorted sediment. Sample 25 has nearly 10% of its volume in the fine fraction but an excellent sorting. The median is not excessively small but this percentage of fine material is sufficient to reduce permeability to values less than 100 meinzers. Sample 22 has about 5% clay— silt. It has good sorting and a median which is not too small. Two factors combine to reduce its permeability. The clay-silt fraction is one and the other is the distri- bution of sizes. The curve shows that a fairly large amount of material fall into each size interval up to very coarse sand. This presents a good chance for interstitial clogging. 97 60 ‘ 50 “ 40 3O 2O 10 0% . m E (\I H In ° N C) C) O (\J I - Ln 0 . . . \o I I . m .: cn o In I I H I I I . m U\ I I I I I V KO (\J LG I C) O O O I—I (\J In . . . o o o v—{ (\I :1- Sample 22: Sample 25: 130 meinzers 81 meinzers Permeability 1.13 .765 Sorting 1.83 2.25 O Median Figure 29.-—Percentile size distribution: Samples 22 and‘25. 98 Figure 30.——Percentile size distribution: Samples 14 and 44. Samples 14 and 44.-~Note the great similarity of the two curves. There is only a slight difference in median size but if the means are noted there is a significant difference. The shape of the two distributions also indicates a possible difference of perme— ability. Sample 14 has a broader curve in the sand sizes whereas sample 44 has a sharper curve. The opportunity for blocking by successively sized particles is less in sample 44. 99 50 r- 40 3O 20 IO 0% ~—---+- -‘ /I I I I P—r I l. E U) E U) LG (I) (\I O H H Ln CD ° ° N O O 0 ° N 08 I ° Ln CD ° ° ° \0 m I I ' ° N 3' 00 H I m\ In I I H I I I I I nI m U\ I I I I I I o \O \O ("\I LG I O C) C) 0 ° 0 O H (\I m o o o . \O o o s . H (\I :I' oo "1 Sample 14: Sample 44: 280 meinzers 860 meinzers Permeability .81 1.74 Sorting 1.59 1.28 0 Median Figure 30.——Percenti1e size distribution: Samples 14 and 44. 100 Figure 31.—-Percentile size distribution: Samples 26 and 30. A. Samples 26 and 30.-—ln these unimodal distri- butions the permeability parallels the median size. Both of these samples have their medians in the medium sand size range and their permeability values are near the average permeability for this size range: 362 meinzers. 101 60 — 50 r 40 30 20 10 0% I I . E Ln E (\I In ’1 n (\l I ~ Ln 0 CD- 0 Q C2 \0 I I * o (\J o . \o O LG I l ”I I :T (D H - (\1 Ln l I l I l l I! \O (\I Ln I C) I I I O H (\I Ln 0 C) O O . . . . ,—I . . . (\J :1' 00 Sample 26: Sample 30: 390 meinzers 330 meinzers Permeability -795 .58 Sorting 1.47 1.94 0 Median Figure 3l.——Percenti1e size distribution: Samples 26 and 30. 102 Figure 32.-—Percentile size distribution: Samples 13 and 34. A. Samples 13 and 34.-—ln comparing these distri— butions and their medians it is apparent that the medians do not parallel the permeability values. The reason might be the sorting value but is more likely the clay-silt per— centage of sample 34. Sample 13 has more of its bulk in the medium sand sizes also. 103 —1 m s (\l E H U\ o (\I ° N O O O 0 KO l. ° m 0 O o o \O O I I . (\I :r 00 H . In I I H I I I I v m U\ I I I I I I \O (\I LG I C) O O C) O .’_‘I (\I Ln 0 o o s o o o r_‘I (\l :- w Sample 13: Sample 24: 360 meinzers 240 meinzers Permeability .365 .805 Sorting 1.94 1.83 0 Median Figure 32.-—Percenti1e size distribution: Samples 13 and 34. 104 Figure 33.--Percentile size distribution: Samples 31 and 41. Samples 31 and 41.--No permeability data was available on these samples but both have medians which fall on either side of the medium—sand—fine—Sand boundary so their permeability is probably not very high. Sample 41 has a distribution similar to sample 30 and an identical median and would probably have a permeability of about 300 meinzers. The permeability of sample 31 should be around 100 meinzers since it has a fairly small median and 5% of clay-silt size material. 105 LG ("\I v-I Ln 0 o (\J O O O . Ln I ' Ln O 0 O o \O (\J I I - (\I :r 00 H \o In I I H I I I I o (M In I I I I I I \O (\I LG I O O O O J O H N m 0 0 o o - - . H m .3 (D Sample 31: Sample 41: No data No data Permeability 1.14 .575 Sorting 2.05 1.93 0 Median Figure 33.-—Percentile size distribution: Samples 31 and 41. 106 Figure 34.-~Percentile size distribution: Sample 5, 19, 16, and 29. .__——.— —— _——._.— — —-— A. Samples 5 and l6.--The permeability of these two samples is not easily explained by noting their statistics. Sample 16 has a larger median, only slightly poorer sorting, and tends to have less clay than sample 5. An examination of the mineralogy of the clay fraction shows that sample 5 has a very crystalline kaolinite as the only clay mineral present, while sample 16 has illite, kaolinite, and vermiculite present. The kaolinite is a non—expandable clay mineral and would impede the progress of fluids less than the illite- vermiculite combination. B. Samples 19 and 29.--These samples illustrate the increasing effect of good sorting with increasing size. Both samples have nearly identical sorting values and the same clay- silt percentages. The difference in median size determines the permeability difference. 107 Sample 5: Sample 16: 310 meinzers 220 meinzers Permeability .61 .825 Sorting 2.05 1.62 0 Median I *1 J 5' Ln E LG (\I L0 0 (\I H (\J O O O ' \O . . LO Q . . . \o O I I 0 0 (\I :1‘ CD r—I - I I I r-I l I l l v m U\ I I I I I I \O (\I LG I O O O O O H (\I m o o o o - - - . H (\I :r 00 Sample 19: Sample 29: 290 meinzers 2100 meinzers Permeability .855 .88 Sorting 2.05 .55 0 Median 40 — 3O — 20 _ 10I— I 0 7.; ' Figure 34.--Percenti1e size distribution: Samples 5, 19; 16 and 29. 108 Figure 35.--Percentile size distribution: Samples 15 and 17. A. Samples 15 and 17.-—No permeability data is available for these samples. Probable values may be inferred as about 100 meinzers for both. They have excellent sorting but have their medians in the fine sand range and the shape of the curves preclude any modification by larger grain sizes. 109 5 7o ‘— 60 ~ AZ 50 - 40 — 30 _. 20 - 10 - 0% I I I I l 1 I. E E LG (\I Ln r—I Ln O (\J - (\I O O 0 ° \O I 0 LO Q o o o \o O l I ' (\J :t' (D H - LG I I, H I I I I m m\ I I I I I I \O (\J Ln l O O O O O H (\1 Ln . . . . . . o r—1 (\1 :1' 00 Sample 17: Sample 15: No data No data Permeability .25 .50 Sorting 2.H9 2.32 6 Median Figure 35.--Percentile size distribution: Samples 15 and 17. 110 Figure 36.-—Percentile size distribution: Samples 11, 24, 12, and 40. Samples 11 and UO.—-The striking effect of a high clay percentage is shown by sample 11. All other indicators of permeability are secondary to a high clay—silt percentage. The permeability of sample 40 can be accounted for by its large median and mean sizes. The heterogeneity of its sizes tend to keep the permeability lower than otherwise expected for the median size. Samples 12 and 24.-—The high clay—silt per- centages again limit the permeability. The means indicate the dominance of small grain size but the medians are not extremely small. lll Sample 11: Sample U0: 0.2 meinzers 220 meinzers Permeability 3.75 1.67 Sorting 2.73 .92 0 Median 50 r 5.34 .265 0 Mean I.- E. Ln OS ("\I Ln 0 . H (\I C) C) O (\I N . . Ln 0 - - . IO N) \o I I - mI .: (D rI I o I I I H I I I I l . m In I I I I I I o V \O (\I LG I O C) O 0 ° 0 H N m ' O 0 KO - o - H (\I :1' CD r—I Sample 12: Sample 2D: 50 r 0.1 meinzer 22 meinzers Permeability \ 3-85 3.75 Sorting 2.83 1.64 0 Median 40 A 5.44 2-52 0 Mean Figure 36.-—Percentile size distribution: Samples ll, 24; 12 and NO. 112 Figure 37.--Percentile size distribution: Samples 32, 39, 36, and 47. A. Samples 32 and 39.——Samp1e 32 is a clay hardpan with an extremely low permeability. Sample 39 has nearly 10% clay and silt but a fairly large median. It probably has a permeability in the 300 meinzer range. B. Samples 36 and M7.—-No permeability data available. The inferred permeability values would be low because of the high percentage of clay and silt. ad 70 113 Sample 32: Sample 39: 0.05 meinzers No data Permeability 3.18 2.70 Sorting 6.80 .88 0 Median /"““‘~—”/ I fl _I E E LG (\I O H In 0 - Ln . (\I O C) O ' (\I (\J | o m C o . . KO m \0 I I . - (\J -:T (D H | o uw I I H I I I I I . m In I I I I I I o V \O (\I LG | O O O 0 ° O H (\J m o o o . \o . . . . H (\I :I' (D H Sample 36: Sample 97: No data No data Permeability 2.41 4.88 Sorting “01' 3.00 2.05 0 Median ‘ 3.965 3.7“ Q Mean 30 20 10 _J 0% Figure 37.-—Percentile size distribution: Samples 32, 39; 36 and 47. THE EFFECT OF SORTING ON PERMEABILITY Table 4 lists the sorting values by increments and their corresponding permeabilities. While there is not a direct relationship between the sorting values and permeabilities a general relationship is evident if the values are grouped. For samples having sorting values less than one Wentworth unit, 29% have a permeability of more than 500 meinzers. However, 12% have a permeability of less than 100 meinzers which excludes them from classi- fication as aquifers. The average permeability for the group is 580 meinzers with a spread from 22 meinzers to 2100 meinzers. Samples having a sorting value of 1—2 Went— worth units have an average permeability of only 457 meinzers but 50% of the samples have a permeability greater than 500 meinzers and only 8.3% have a permeability less than 100 meinzers. For samples having a sorting value over two the permeability values are less than 100 meinzers with few exceptions, and for sorting values of three and more the permeabilities are less than 10 in all the samples tested. The conclusion may be drawn from the above discussion that sorting is a limit to permeability if it is too good and also when it is very poor. If sorting alone is considered, however, it would not be valid to place limits relating to 114 115 omm HN.H m QQMH Hm.H mH OMH MH.H mm 03H m:.H mm mthCHmE OOH can» mmmH oom om.H mm mmHUHHHomoEpmd m>mc Rm.m .mpmNchE oom III :H.H Hm swap pmpmmpm mmfipfififipmmELma m>mc aom omH os.a am mhmuchE 5m: thHHnmmEhma mwwhmw< 0mm 5m.H o: m I H n m zm mm.H m: 0mm Hw.H m: 0mm 3N.H a: mpmNCHmE com am.H m: am: HmpOQ go a cam Ho. m com Hm. m 0mm :m. m 02m mm. OH 0mm mm. MH 0mm Hm. :H III om. mH mpmNchE OOH omm mm. ea swap mmmH no moHpHHHpmmEan m>mz III mm. NH &~.HH ocm mpmNchE com cmcu pmpmmmw 0mm mm. mH mmfipfiafinmmspma m>mg o.a swap mama m 0mm ms. om wcH>ms moHdEmm mo amm co: mm. Hm mm mu. :m me I mpHHHommEpmo owmgm>< Hm mm. mm 0 0mm mm. mm o.H m OQOH Hm. mm QQHN mm. mm omm mm. om ozm mow. 3m mpmNchE III mam. H: spHHHnmmspmm om .oz mHQEmm .spaafipmmsnma .m> mcfipaomII.z mamae 116 om. mm.m HH 30H mho> mpHHHmeEpwm OH. mm.m NH +m n m mo. wa.m mm III mm.: m: Ram III a:.m mm owm mm.m mm mCOHmSHocoo smug ow mono pcmHOHMQSmmH III H:.m mm m I m u m mw om.m mm III o~.m mm III o:.m m: szHHnmoEpom om .oz mHQEmm Umscfipcooll.: mqm¢8 4117 ‘.062nn. o . 4.75 '- 4.50 _ 4.25 1 _ 4.00 _ ‘ 3.75 - 5.50 —" silt 3.25 2.6 1.6 —I ..4 l l I .47 ‘ flsoa'rme In. fluepuu .11 I!— very very coarse medium fine very r-fine coarse aand sand sand fine gravel. sand ,aand _ .35 " 58 9 O . 3 53 _. .45 .56 _ . 42 06" (I ~ '5 I (I45 _ 4o 8 .44 Of" 18 O ‘ .zz. “ I.28 ' .7 22...?»1 -> I.I48 _ 29 . .19 26.’ 16 14 “’12 as 27 {#50 .5I I. __ 2E. ’15 30"II:1 ['15 4 2 1 .5 9’19. 15 17 .125 I L I I!;1__Jh. I .2 .Figure 38.~~0 Sorting Compared to a Md.‘ .105 .1.0 -05 0 ' .5 ‘1.0 1.5 2.0 2.5 5.0 5.5 4.0 5.0 6.0 7.0 8.0 fill 118 permeability. Regardless of how good the sorting may be, if the median size is not large enough the material will be a poor aquifer. The shape of the percentage distribu- tion curve will tell at a glance whether the bulk of the material is sufficiently large if combined, with a good sorting value, to give a chance of good permeability. The Effect of Sphericity on Permeability A comparison of sphericity values with permeabili- ties of the same samples shows no pattern of correlation. There is a great difference in permeability values for samples having similar sphericity values. Table 5, which follows, shows the sphericity values arranged according to equal intervals and the corresponding permeability values. The average permeability value for each group is given. The average permeability in this case is not a true indica— tion of the spread within the group. Little information can be inferred from the sphericity data so this parameter is assumed to be of only minor influence with regard to permeability. The Effect of Roundness on Permeability Roundness values are shown in Table 6 (page 121) with the corresponding permeability values. The roundness values correlate with the permeability values so poorly that conclusions can not be drawn from them. Both the sphericity and roundness values indicate that the glacial 119 Ill‘lllllllllllllllll"lllllllIll'llllllll[IIIIIll'II'IllII'I'Il'l'5ll'5l'l'5lll'l 0mm mmo. :: III :00. mm III mmo. mm omm 3mm. om mamNchE mam u moHHHomoELmd mwmpo>< 0mm mmo. mm mam. I one. I synonymsam am :50. mm coma cam. ma III mom. ma ozm mam. as am mam. m: mumNcHoE mmm u muHHHpmmano mmmam>¢ omm mmm. 0: 0mm. I mmo. I spHOHpmeam cam 0mm. mm com 020. mm OQOH mmw. um oom Hmo. m: QMH mHm. um mpmmcHoE mmm u mpHHHQMmELoQ mwmho>¢ co: 2mm. Hm mmm. I mam. I apHOHnmzam H.o mam. NH 0mm smm. m mpHHHnmmeumm szOHhmndm .oz oHQEmm .mpflafinmmsnma .m> sSHOHsmzamII.m mamas 120 mo.o Ham. mm mHoNchE OHN u muHHHnmmEamo owmpm>< OOHm Haw. mm om». I mmw. I apneanmnam omH amp. mm QHm mmw. 5 III mmw. m: mo mHm. mm omm MHN. mH mnmNchE mom I zuHHHommEpmd mwmam>¢ III mHm. NH mma. I ooa. I apHOHnmegm 0mm How. on 0:0 How. OH OHm HHN. m II mac. 5: 0mm owe. me III mmm. H: cam owe. am II mum. Hm mm mam. am 03H mmo. mm omw moo. om owm owe. 2H com 0mm. ma mamwchE mzm u zuHHHpmmEme.ommam>¢ m.o mam. HH cop. I mam. I apHOHnmgam 0mm one. 0 0H0 :wo. 5 III mum. : III mam. m spHHHQmIEnmm announmcam .oz maaemm UmchpcooII.m mqm< QQQH mam. mm omm. I com. I mmmcecsom 0mm mmm. om owm Ham. :H com 3mm. mH 0mm Hmm. m OH.m 2mm. 5 QHm omm. m III ozm. : III mmm. a: :m mmm. m: 00: mwm. Hm mpmNchE mmH u szHHnmoano mwmpm>< H.o mmm. mH com. I 0mm. u mmmcoczom com mam. m III mom. m mpHHHommanm mmmcocsom .oz mHoEmm .mpHHHnmoELod .m> mmmCchomII.w mqm¢ 00Hm 0::. 0m + 00:. u mmmcoczom mm 0m:. :m QMH mm:. mm 00m :0:. 0H com Hmm. a: 0mm m0m. 0: III mmm. mm mm 0am. 0m 0MH 00m. 5m 0:m mom. :m III mmm. Hm 2 0mm mum. Om m oom 30m. mm 00m 05m. 0m Hm mam. mm 0:H 00m. mm maoNchE mom I HpHHHQonme owmao>¢ 00MH :wm. 0H 00:. I 0mm. n mmmcncsom III 00m. 0H 0mm 00m. 0H III :0m. mH m.0 Hom. HH 0:0 mmm. 0H 0mm mmm. 0 III mam. m szHHommELom mmmcvczom .oz mHQEmm uochpcoolI.0 mqm¢e 123 sediments in the samples are quite spherical (approaching equal dimensions) and are rather poorly rounded. An exam- ination of these sands shows this to be true. The particles have very sharp corners in many instances and yet are not strongly elongate. The Effect of Mdfl on Permeability The analysis of the median size in the sediments involved shows a number of interrelationships. Table 7 on page 125 shows the medians arranged according to increments in Wentworth grades. The following points are brought out: 1. 90% of the samples have their medians in the sand size range from 0 - 3 0. 2. 23% have their median size in the coarse sand range (.5—1.0 mm.). 3. 96% have their median size in the medium sand range (.25—.5 mm.). 4. 23% have their median size in the fine sand range (.l25-.25 mm.). 5. 76% have a sorting value of less than two Wentworth units. 6. 20% have a sorting value of less than one Wentworth unit. From the preceding several conclusions may be drawn. The predominant grain size in the sediments falls into the medium sand size range. Nearly half of the sediments have their medians in this range. This, with the related spread of medians on either side indicate that these sediments are mostly sands and not tills or clays. Since they are located in and near a pre-glacial drainage channel one can expect 12b to find this true. These sands are very well sorted as shown by the percentages having sorting values of less than two Wentworth units. Twenty per cent of the samples have sorting values of less than one Wentworth unit. How the median size is related to the permeability is shown in Table 7 (page 125). Of the samples having their median size in the coarse sand size range 70% have a permeability greater than 500 meinzers and 30% have a permeability of less than 500 meinzers. The average permeability for this group is 745 meinzers with a range from 2100 to 220 meinzers. 0f the samples having a median in the medium sand size range 28% have a permeability greater than 500 meinzers. Twenty—two per cent of these samples have a permeability of less than 200 meinzers and 50% have permeabilities between 500 and 200 meinzers. Their average permeability is 362 meinzers with a spread from 860 meinzers to 22 meinzers. In the fine sand range only 16% of the samples have permeabilities of more than 200 meinzers, 67% have permeabilities of less than 100 meinzers, and 16% have permeabilities between 100 and 200 meinzers. Their average permeability is 98 meinzers with a spread from 0.1 meinzer to 310 meinzers. A permeability of less than 100 meinzers precludes the classification of the unit as any kind of an aquifer. From the preceding discussion it is evident that the median size of a sediment is closely related to the actual 125 :0.H mm :m 00.H :m m: 00.H omH mm mpmNchE 00m A 00> m0.H 0mH mm I m0.H 0mm 0H mamacflms com A amm mw.H oam am 00.H 00m 0 mLoNcHoE mom I mpHHHnmmEme owmpm>4 00.H 00m :H :0.H 0mm 0m UQmm Eszmz :0.H 00m MH .55 :\H I m\H ma.H so: am 0 m I H Hm.H 0M0 0 :0.H 0:0 0H 5:.H 00m 0m mm.H omm m: mH.H 0mm 0m 0m.H 000 :: 00.H 000H mm mammcflms oom AIaooa mm. III am mamNchE 000 I 00w m0. 0mm 0: maoNchE m:m u mpHHHnmoELoQ owmpo>< :0. 00m 0H comm mmpmoo :0. 00m 0 .EE m\H I H H:. 000 mm 0 H I 0 00. 000 0: :0. 0H0 m 00. 0:0 0: 0H. 00mH 0H mm. 00Hm 0m Hm>mm0 mw.0I III 0: .EE : I H H0.0I 00m mm a m- I 0 am.:- III mm a a: apfiafinmmssmm .oz maasmm .cmHUoE 0 .m> mpHHHQmE®L®mII.N mqmdB 126 .EE 0\H pHHm 0» 0cmm mch 00.0 00.0 00 .> I 0.0 00.0 III 00 .EE 0\H I 0H\m 020m mch 00.0 0.0 HH 0 m I 0.0 00.0 H.0 0H mpmNchE 000 A 00: onNCHmE 000 M 0:02 00.0 III Hm mamNcHoE 00 u szHHnmmEamQ mwmam>< 00.0 III 0: 02mm mch 00.0 III 0H .88 0H\m I :\H 00.0 00 00 0 0.0 I 0 00.0 H0 00 0H.0 0:H m0 0:.0 000 0H 00.0 0Hm 0 a as mafiaanamEpmm .oz magsmm UmSCHpCOOII.0 mqm