UNIVERSITY OF OKLAHOMA GRADUATE COLLEGE STRATIGRAPHY AND RESERVOIR CHARACTERISTICS OF THE DESMOINESIAN GRANITE WASH (MARMATON GROUP), SOUTHERN

Transcription

1 UNIVERSITY OF OKLAHOMA GRADUATE COLLEGE STRATIGRAPHY AND RESERVOIR CHARACTERISTICS OF THE DESMOINESIAN GRANITE WASH (MARMATON GROUP), SOUTHERN ANADARKO BASIN A THESIS SUBMITTED TO THE GRADUATE FACULTY in partial fulfillment of the requirements for the Degree of MASTER OF SCIENCE By ALYSSA MARIANA KARIS Norman, Oklahoma 2015

2

3 Copyright by ALYSSA MARIANA KARIS 2015 All Rights Reserved.

4 This thesis is dedicated to my younger cousins, nieces and nephews. Let this be an inspiration to all of you to pursue your dreams. No matter how impractical they may appear, it is possible to achieve them and I will be there every step of the way cheering you on. I love you all so very much.

5 Acknowledgements This research was funded through the sponsors of the Reservoir Characterization and Modeling Laboratory and the Granite Wash Consortium: Chesapeake Energy, Devon Energy, QEP Resources, and SM Energy. I thank all the sponsors for providing data for this study. I would like to thank my advisor, Matthew Pranter, for his extensive guidance, advice, and assistance throughout this whole process. I would like to express my appreciation for my committee members, Doug Elmore and Deepak Devegowda, for answering questions and reviewing and editing my thesis. I would also like to thank John Mitchell for answering questions and providing help for this project. I would like to express my gratitude to Suriamin who guided me throughout the petrophysical analysis and thesis process. The following companies provided invaluable software that allowed me to complete this project: Schlumberger (modeling), IHS (subsurface interpretation), Paradigm (well-log petrophysical analysis), and LR Senergy (cluster analysis). I would like to also thank my family, Mom, Dad, and Tim, for always supporting and encouraging any project I take on. I could not have accomplished all that I have without you and I love you. Last but not least, I would like to thank my fellow graduate students in no particular order: Colton Birch, Elizabeth Gergurich, Suriamin, David Tilghman, Marcus Cahoj, Joseph Snyder, Gabriel Machado, Sumit Verma, Carolina Mayorga, Javier Tellez, Josh Hardisty, Anna Turnini, and Brandon Swain. Your friendship has been a great source of delight and support for me over the last two years. iv

6 Table of Contents Acknowledgements... iv List of Tables... vi List of Figures... vii Abstract... viii Introduction... 1 Tectonic and Stratigraphic Setting... 7 Petrophysical Analysis of Lithology Electrofacies: Log-based Lithology Estimation Structural and Stratigraphic Framework Spatial Distribution of Lithology Conclusions References Appendix A: Core Description Appendix B: XRF Mechanics Appendix C: Normalization Histograms Appendix D: Type Log Appendix E: Regional Cross sections, Structure Maps, and Isopachs Appendix F: Model Area Cross sections, Structure Maps, and Isopachs Appendix G: Vertical and Horizontal Variograms Appendix H: Model Results Appendix I: Petrophysical Equations v

7 List of Tables Table 1: Data distributions for k-means algorithm Table 2: Point comparisons of actual lithologies versus predicted lithologies Table 3: Well-logs value cut-offs for each lithology Table 4: Variogram inputs for SIS lithology modeling Table 5: Variogram inputs for SGS effective porosity and water saturation models Table 6: Lithology output percentages Table 7: Arithmetic means for effective porosity and water saturation vi

8 List of Figures Figure 1: Tectonic Provinces of Oklahoma and Texas... 2 Figure 2: Study Area Maps... 6 Figure 3: Pennsylvanian Paleogeography Map... 8 Figure 4: Anadarko Basin Cross Section... 9 Figure 5: Stratigraphic column for the Oklahoma Granite Wash Figure 6: Environment of Deposition Figure 7: Mayfield 1-34 Type Log Figure 8: Core photographs, wire-line signatures, and permeability-porosity cross plot Figure 9: Core Chemostratigraphy Figure 10: Cluster Analysis Cross plots Figure 11: Lithology Estimations Figure 12: A-A' Regional Cross Section Figure 13: B-B' Regional Cross Section Figure 14: Structure and Isopach Maps for the Marmaton Group Figure 15: Marmaton Group Model Grid Figure 16: Marmaton Group Model Structure Figure 17: Vertical Lithology Proportion Curve Figure 18: SIS Lithology Model Figure 19: Combined Conglomerate and Sandstone Percentage Map Figure 20: SGS Effective Porosity and Water Saturation Models Figure 21: Comparison the Three Models vii

9 Abstract The Desmoinesian Granite Wash, specifically the Marmaton Group, is a hydrocarbon-bearing interval within the Anadarko Basin of Oklahoma and Texas that is composed of clastic and carbonate sediments derived primarily from the Amarillo- Wichita Uplift. The Marmaton Group, located in Beckham County, Oklahoma and Wheeler County, Texas, includes a series of vertically stacked conglomerates and tightgas sandstones and shales that exhibit a complex stratigraphic architecture, highly variable lithologies, and correspondingly heterogeneous reservoir properties. The stratigraphic and reservoir characteristics of the Marmaton Group, are established based on cores, x-ray fluorescence (XRF) measurements, and well-log signatures. The Marmaton Group in the southern Anadarko Basin contains interbedded arkosic sandstones and conglomerates that thin laterally into shales to the north (basinward). At least four regional, correlatable flooding surfaces (and associated organic-rich shales) subdivide the Marmaton Group and are thought to be self-sourcing in this liquids-rich interval. Porosity in this interval varies from 2-18% with low permeabilities on the order of 10 µd. Proximal to the Amarillo-Wichita Uplift, the Marmaton Group is highly lithologically heterogeneous. XRF analyses of cored intervals show that elemental concentrations vary stratigraphically in conjunction with lithology. Characteristic welllog signatures correspond to different intervals and can be correlated laterally through the study area. Cluster analysis implemented on well-log data resulted in a 63% correlation to the Mayfield 1-34 core description but achieved low correlations for the Mayfield 1-2 (0%) and Sage 1-34H (53%). Well-log cutoffs performed on well-log data viii

10 have a 74% correspondence rate to Mayfield 1-34 core description. Overall the well-log cutoff lithologies provides an approximation of lithologies in non-cored: 62% sandstone, 23% conglomerate, and 15% shale. A compiled lithology model of the Marmaton Group displays spatial patterns by zone constrained to the vertical lithology proportion trend, vertical variograms, horizontal variograms, and lithology percentages. Using the lithological trends as an input, effective porosity and water saturation show that conglomerates on average have a higher effective porosity (by 1%) lower water saturation (by 1%) throughout the Marmaton Group. ix

11 Introduction Sedimentary deposits that comprise the Granite Wash play of Oklahoma and Texas are present in the southern Anadarko Basin (Figure 1). The source area for the Granite Wash, the Amarillo-Wichita uplift, forms the southwestern margin of the Anadarko Basin and has produced a complex tectonic setting with numerous faults and folds. Formations vary stratigraphically and laterally due to changes in environment of deposition and ongoing tectonism. Properties of reservoir rocks, such as porosity, permeability, water saturation, and pressure, are challenging to characterize and map spatially. This, historically, has made original oil- and gas-in-place calculations and recovery from vertical wells difficult (Mitchell, 2011). Recent trends in horizontal drilling have increased the productivity of formations ranging from Atokan to Virgilian in age. Immediately northeast of the Amarillo-Wichita uplift in the deepest part of the Anadarko Basin (the proximal play area) are the thickest intervals of Granite Wash alluvial deposits (conglomerates and sandstones). These deposits decrease in grain size and become thinner to the northeast (basinward) (Mitchell, 2011). The proximal play area, which encompasses the study area, consists of conglomerate (arkosic and lithic), sandstone (arkosic and lithic), and mudstones as, and the Desmoinesian Marmaton Group is a major producing interval. Oil, gas, and condensates are produced from the Marmaton Group and other formations and have a total cumulative production, as of February 2011, of 222,458 BBL (35,368 m3) of oil and 2.2 BCFE (6.22 x 107 m3) of gas (Mitchell, 2011). The distal play area, near the middle of the basin as defined by Mitchell (2011), has reservoirs that are predominately arkosic 1

12 OTTAWA WOODS MAJOR ALFALFA GRANT GARFIELD NOBLE KAY WASHINGTON OSAGE 36 BLAINE KINGFISHER LOGAN PAYNE PAWNEE TULSA NOWATA ROGERS CRAIG DELAWARE WAGONER CREEK ADAIR LINCOLN OKFUSKEE SEMINOLE CADDO HASKELL 35 HUGHES LATIMER PITTSBURG COMANCHE COTTON CANADIAN GRADY STEPHENS JEFFERSON OKLAHOMA LOVE POTTAWATOMIE GARVIN MURRAY CARTER PONTOTOC COAL JOHNSTON MARSHALL BRYAN ATOKA CHOCTAW MAYES BRAVO DOME McINTOSH DALHART BASIN CIMARRON ARCH ANADARKO SHELF / RAMP CIMARRON CHEROKEE PLATFORM HARPER BEAVER TEXAS NEMAHA UPLIFT OZARK UPLIFT WOODWARD ELLIS 36 CHEROKEE DEWEY OKMULGEE MUSKOGEE ROGER MILLS ANADARKO BASIN SEQUOYAH CUSTER BECKHAM ARKOMA BASIN CLEVELAND WASHITA 35 LE FLORE McCLAIN KIOWA HARMON GREER ARBUCKLE UPLIFT PUSHMATAHA HOLLIS-HARDEMAN OUACHITA UPLIFT UPLIFT PARMER DALLAM HARTLEY OLDHAM DEAFSMITH SHERMAN MOORE HUTCHINSON ROBERTS POTTER CARSON GRAY RANDALL HANSFORD ARMSTRONG OCHILTREE LIPSCOMB Study Area HEMPHILL View Mountain WHEELER Fault System DONLEY COLLINGSWORTH AMARILLO WICHITA PALO DURO BASIN Meers Fault CASTRO SWISHER BRISCOE HALL CHILDRESS JACKSON ARDMORE TILLMAN McCURTAIN MARIETTA BASIN BASIN 99 BASIN United States Oklahoma N 50 mi 80 km Figure 1. Map of tectonic provinces for Oklahoma. The study area is shown. (modified from Johnson and Luza, 2008; Northcutt and Campbell, 1995; Campbell, et al., 1988; Dutton, 1984; LoCriccho, 2012, McConnell, 1989) 2

13 and dolomitic sandstones, which mostly produce gas and condensate. The cumulative production totals as of February 2011 were 119,027 BBL (1,8924 m3) of oil and 28 BCFE (7.93 x 108 m3) of gas (Mitchell, 2011). Prior research on the Granite Wash has focused on the stratigraphy and various lithologies to better understand the environments of deposition. Brown (1979) studying Pennsylvanian deltaic sandstone facies in the Mid-Continent region using electric logs and models and Mitchell (2011) in the Texas Panhandle and western Oklahoma, have interpreted the depositional environments of the Granite Wash interval as alluvial fans, deltaic fans, proximal turbidites, and debris flows. Alluvial fans require an uplift that sheds a large amount of debris (weathered rocks) down the slope and are deposited close to the Amarillo-Wichita uplift (McGowen, 1971; Brown, 1979; Mitchell, 2011). This creates a poorly sorted deposit containing grain sizes from gravel to mud. McGowen (1971) studied modern day alluvial fans and fan deltas to produce depositional models and Brown (1979) found deposits nearer to the uplift (source) will contain more gravel, while downdip will be predominately clay-sized grains with a few large clasts. McGowen (1971) explained that fan deltas are created through the same process as alluvial fans but prograde into lacustrine or marine environments as the water level rises. They typically have sands interbedded with floodplain deposits. Brown (1979) determined the morphology of the fan deltas juxtaposed to the Mountain View Fault to be elongated to lobate deposits. This indicates that the marine environment of deposition had low energy (Brown, 1979). Duggins (2013) studied facies architecture and sequence stratigraphy of the Marmaton Group in the Texas Panhandle and western Oklahoma using wire-line logs, core, and magnetic susceptibility. The highest reservoir 3

14 quality found in that study was in the upper fan delta apron sandstones, proximal to the Amarillo-Wichita uplift. These substantial fan delta apron deposits had some of the highest porosity and permeability (Duggins, 2013). The Marmaton Group has been correlated in the southern Anadarko Basin using four black, radioactive marine shales (and associated flooding surface). These marker beds thicken towards the basin and represent marine transgressions as well as separate fan delta sandstone deposits (Sahl, 1970; Dutton, 1984). These shales possess an easily identifiable spike in gamma-ray (GR) and neutron porosity (NPHI) readings and a reduction in deep resistivity (ILD) and density porosity (DPHI) readings (Mitchell, 2011). Determining lithologies from basic well log analysis proves to be challenging and oftentimes impossible. Principal component analysis (PCA) combined with cluster analysis aims to identify the important variables in a problem, or in this case, estimating lithologies in non-cored wells. This concept is also known as electrofacies (Serra and Abbott, 1980), which is the method of relating rock types, not genesis origins, to well log signatures (Doveton, 1994). Esbensen and Journel (1992) explored several methods for scrutinizing geologic data using multivariate statistical approaches. PCA reduces the dimensions of the data without making substantial statistical assumptions, therefore keeping the character of the data intact. It takes data plotted in mathematical space (theoretically infinite number of axes) and calculates the first and second eigenvectors (major and 1st minor axes) from the correlation matrices (Doveton, 1994). The reduction of data, and therefore space, allows for faster computing. Once PCA is applied, cluster analysis could be used to classify the data points (Esbensen and Journel, 1992; Doveton, 1994). Cluster analysis using algorithms to minimize data point distances to a center 4

15 point, thus it is crucial to identify, and when necessary delete, outliers in order to create clusters that geologically make sense. Although this technique is actually unsupervised it is best suited for previously classified lithologies (Esbensen and Journel, 1992; Doveton, 1994). Kaźmierczuk and Jarzyna (2006) used principal component analysis (PCA) and cluster analysis on log data to classify three reservoir lithologies in the Carpathian Foredeep. PCA reduced the dimensionality of their data set from 15 primary well logs to two principle components and used a variance-based cluster analysis approach (Kaźmierczuk and Jarzyna, 2006). This study reported PCA in conjunction with cluster analysis to be efficient, use less space, and produce lithologies similar to those interpreted from qualitative means. Employing PCA along with k-means cluster analysis guided by core descriptions would help predict lithology in uncored wells. This study explores the complexities of the Marmaton Group throughout an area of 728 mi2 (1.89 x 103 km2) that spans Beckham and Roger Mills Counties, Oklahoma and Wheeler County, Texas. Data from 430 wells with wire-line logs and three cored wells (Figure 2) were used to explore key lithologies and their petrophysical properties. The main objectives are to 1) determine the key lithologies, petrophysical properties, and unique well-log signatures or values associated with certain lithologies or petrophysical properties; 2) Define the structural and stratigraphic framework of the Marmaton Group throughout the region and; 3) map the spatial distribution of lithology, porosity, and water saturation. 5

16 A A B N TX OK Wheeler Roger Mills Model Area B Beckham A Wells with digital and/or raster logs 4 mi 6.4 km B 11N 26W 11N 25W N 10N 26W 10N 25W Wells with digital logs 3 mi 4.8 km Figure 2: Study area and model area for this study. A) The study spans Wheeler County, Texas and Roger Mills and Beckham Counties, Oklahoma. There are 430 wells included in this study. Lines A-A and B-B are regional cross-sections of Figures 12 and 13 respectively. B) The model area shows only wells that contain GR, DPHI, NPHI, and ILD digital well logs. The Mayfield 1-34 (star, API ) core was described in this study and the cored wells with published descriptions are the Sage 1-34H (triangle, API ) and the Mayfield 1-2 (square, API ). 6

17 Tectonic and Stratigraphic Setting The main tectonic features associated with the Granite Wash area in western Oklahoma and the Texas panhandle are the Wichita-Amarillo uplift and the Anadarko Basin. Basin formation, uplift, and deposition of the Granite Wash in southwestern Oklahoma occurred in multiple phases. The Amarillo-Wichita uplift began forming in the early Pennsylvanian (~318 Ma) (Figure 3) as a result of the Ouachita-Marathon orogeny (Wickham, 1978; Kluth and Coney, 1981; McConnell, 1989). The OuachitaMarathon orogeny was a continent-continent style collision between the North and South American plates (Graham et al., 1975) and reactivated Cambrian-aged Southern Oklahoma aulacogen faults in the Amarillo-Wichita uplift and Anadarko Basin (Kluth and Coney, 1981). Large-scale, basement-involved reverse faulting during the Morrowan created the asymmetry of the Anadarko Basin and formed the deepest part of the basin adjacent to the front of the uplift (Figure 4) (Eddleman, 1961). Reverse faults and inverted reverse faults associated with the uplift strike to the northwest and exhibit a total vertical displacement of 40,000 ft (12,200 m) and left-lateral displacements of mi (12-26 km) (McConnell, 1989; Johnson, 1989). The Amarillo-Wichita uplift and Anadarko Basin subsidence did not produce igneous material due to the distance from the Ouachita-Marathon orogeny. However, the Anadarko Basin accumulated up to 18,000 ft (5,490 m) of sediment throughout the Pennsylvanian (Johnson, 1989). Cambrian-aged granite was exposed on the AmarilloWichita uplift and served as the parent rock for the sediments shed during the Desmoinesian (Moore, 1979; Johnson, 1989). The Marmaton Group (Figure 5), the main focus of this study, contains interbedded arkosic sandstones and conglomerates 7

18 N Eq Amarillo-Wichita Uplift Study Area Ouachitas Figure 3: Middle Pennsylvanian (308 Ma) paleogeographic map (modified from Blakey, 2013). The building of the Amarillo-Wichita uplift started prior in the Early Pennsylvanian and continues through to the end of the Pennsylvanian. Even though the mountain range is still undergoing positive uplift, there is still a stark contrast between the elevation between the mountains and Anadarko Basin. The study area is located just south of the Equator at this time and lies directly infront of the Mountain-View fault system in the deepest part of the Anadarko Basin. 0 0 mi km

19 A B WICHITA MTNS. ANDARKO BASIN Permian Sea Level 10,000 ft 3,000 m Pennsylvanian Mississippian Sil Dev. Precambrian 10,000ft 3,000 m 20,000 ft 6,000 m 30,000 ft 9,000 m Late Cambrian-Ordovician Early Middle Cambrian 0 0 mi km ,000 ft 12,000 m MAJOR LITHOLOGIES B Sandstone and shale Salt, anhydrite, and shale Conglomoerate ( granite wash ) Black shale Cimarron Arch Nemaha Uplift Limestone and dolomite Rhyolite, granite and gabbro A Shale, limestone and sandstone Granite and rhyolite Wichita Mts. Figure 4: SE-NW structural cross section of the Anadarko Basin. Adjacent to the Amarillo-Wichita uplift is the deepest part of the basin, producing a significant asymmetry. Accomodation space juxtaposed to the uplift allowed for a large sediment accumulation (after Johnson, 1989; Dutton and Garnett, 19889; Pippin, 1970). 9

20 System Series Group Unit Virgilian Shawnee /Cisco Shawnee Wash Heeber Sh Pennsylvanian Missourian Desmoinesian Atokan Morrowan Douglas /Cisco Lansing /Hoxbar Kansas City /Hoxbar Marmaton Cherokee Atoka Morrow Haskell Sh Tonkawa Ss Cottage Grove Wash Hoxbar Wash /Sh Hogshooter Wash Checkerboard Wash Cleveland Wash Marmaton Wash Upper Skinner Sh Upper Skinner Wash Lower Skinner Sh Lower Skinner Wash Redfork Ss& Sh AtokaWash 13 Finger Ls Upper Morrow Squawbelly Ls Lower Morrow Marmaton Wash Marmaton A Marmaton B Marmaton C Marmaton D Marmaton E Marmaton F Figure 5: Stratigraphic column for the Oklahoma Granite Wash (modified from Mitchell, 2011). The Marmaton zones are based on shale breaks and regional flooding surfaces throughout the study area. The Desmoinesian Granite Wash has different nomenclature in different states and petroleum companies. In order to draw comparisons to other literature, the following is a guide for the nomenclature adopted in this study. Seven intervals of the Marmaton Group has been divided into (Wash, A-F) and their equivalents are as follows: Marmaton B = Carr, Marmaton C = Caldwell/Britt, Marmaton D = Granite Wash A, Marmaton E = Granite Wash B, and Marmaton F = Granite Wash C. These intervals are used in this study to better examine stratigraphic, lateral, stratigraphic, and petrophysical trends in the Marmaton Group. 10

21 proximal to the mountain front that thin into a sediment-starved interval comprised of shales in the northeastern direction (Moore, 1979). Pennsylvanian deposits adjacent to the Mountain View Fault (Figure 1) are indicative of a transitional environment that includes both terrestrial and marine deposits (Edwards, 1959). The Midcontinent Pennsylvanian epi-continental sea covered the basin during the Desmoinesian. The large contrast of high relief due to the Amarillo-Wichita uplift and the deep axis of Anadarko Basin created a gravel-rich alluvial fan environment (Figure 6). Fluctuations in sea-level created cyclic deposits that are comprised of sandstone and conglomerate deposited during relative sea-level lowstand capped by transgressive limestones and shales (Johnson, 1989; Mitchell, 2011). 11

22 MOUNTAIN FED ALLUVIAL FANS & FAN DELTAS Amarillo-Wichita uplift NARROW ZONE LITTORAL Surface Plume HUMMOCKY LOBES & SPLAYS TALUS AVALANCHING INERTIA FLOW TURBIDITY FLOW BASIN PLAIN Figure 6: Schematic illustration, modified from Reading and Richards s 1994 example for multiple-source gravel-rich ramp, showing the depositional environment for the Granite Wash in the study area. The sediment source is the Amarillo-Wichita uplift. The sediment forms alluvial fans on the steep shelf, allowing the transport of coarse material into deeper water. 12

23 Petrophysical Analysis of Lithology The key lithologies of the Marmaton Group were evaluated through a detailed analysis of core from the Mayfield 1-34 well (Figure 7), XRF measurements on the same core, and open-hole logs from 430 wells. The Mayfield 1-34 core is in the Mayfield Northeast field, Beckham County, Oklahoma (Figure 2). Three cored intervals for this well are housed in the Oklahoma Petroleum Information Center (OPIC): 24 ft (7.3 m) in the Missourian interval, 41 ft (12.5 m) in the Desmoinesian interval, and 13 ft (4 m) in the Atokan interval. This study focuses solely on the Desmoinesian interval (Marmaton Group), thus the 41 ft (12.5 m) of core from the Marmaton F was analyzed. Detailed lithology descriptions (Figure 8, Appendix A) document bedding, color, grain size, and sedimentary structures. Comparison of Mayfield 1-34 well logs to the core description was used to delineate well-log signatures for various lithologies. Core descriptions and porosity and permeability data from the Mayfield 1-2 and Sage 1-34H wells were also used (Figures 2 and 8). The three major lithologies identified in the cored interval are conglomerate, sandstone, and shale. Two types of conglomerate exist: light grey clasts of volcanic and plutonic rock fragments of variable size with a closed framework (orthoconglomerate) and clasts in a dark grey, mudstone-rich matrix (paraconglomerate). They exhibit different gamma ray (GR), neutron porosity (NPHI), density porosity (DPHI), and deep resistivity (ILD) signatures (Figure 8). Sandstones range in color from beige to gray. They are often paired with mudstone to form lenticular, wavy, and flaser bedding. For the sandstone and conglomerate lithologies, fining upward and coarsening upward sequences exist throughout the cored interval. A thick interval of black shale that 13

24 0 GR Ft DPHI NPHI ILD Marmaton Wash Marmaton A Marmaton B Marmaton C Marmaton D Marmaton E Marmaton F Figure 7: Type log from the Mayfield 1-34 (API ) wire-line logs with formation tops (for location refer to Figure 2). The green box encapsulates the 40 ft (12.2 m) of core described and used in this study. 14

25 A B E Permeability (md) 1 in 2.54 cm NPHI 1 in GR DPHI ILD cm NPHI GR DPHI ILD D C 1 in 2.54 cm NPHI 1 in GR DPHI ILD NPHI cm GR DPHI ILD Conglomerate Sandstone Shale Porosity (%) Figure 8: The Mayfield 1-34 core pictures, associated well log signatures, and porosity-permeability plot. A) Typical closed-framework conglomerate found throughout the core and has large variation of grain size. Low GR and ILD. DPHI is consistently higher in value than NPHI. B) The dark grey, matrix-supported conglomerate, has higher GR, ILD compared to the closed-framework conglomerate. The NPHI values cross-over with the DPHI values frequently. In well-log curves, this open framework conglomerate appears to be closer to a sandstone signature. C) Massive medium grained sandstone with fine, dark grey layers of silt throughout. D) Black shale break in the cored interval that has a very high GR signature and NPHI values greater than DPHI values. Core photographs courtesy of OPIC. E) Porosity-permeability cross plot for the Mayfield 1-2, Mayfield 1-34, and Sage 1-34H wells. Conglomerate has a cluster of data points with porosity between 4 and 8% with permeabilities of md. Most sandstone data points lie between 5 and 10% porosity with greatly varying permeability (between 3x10-4 and 8 md). Shale, as expected have low, porosities (4-6%) and permeabilities ( md) 15

26 corresponds to a large spike in GR readings starts at 13,668 ft (4,166 m) and ends at 13,674 ft (4,167.8 m). The core and log signatures show elements of a transitional environment that fluctuates between terrestrial and shallow marine. Conglomerates with closed framework (clast supported) suggest a fluvial/alluvial deposit because the higher energy could transport clasts that large. On the other hand, the open framework conglomerates are likely formed by turbidity or debris flows and mass wasting events. This would account for the assortment of grain sizes and the high mudstone content. Sandstones with wavy, flaser, and lenticular bedding imply a tidal controlled environment of deposition. The switch from terrestrial to marine could be accounted for by either changes in relative sea level or changes in sediment supply. A variety of shales are seen throughout the well logs. Intervals with increased gamma ray and high NPHI can represent fluvial flood plain, estuarine, or deep water deposits. The marine flooding surfaces defined by literature and in this work contain very high GR (up to 250 API) and NPHI (up to 0.23 v/v) readings. The dramatic increase in well log signatures, overall fining upward sequences leading to the shale, and thickness of shale suggest that these specific shales are transgressive flooding surfaces. The coarsening upward sequence from shale to sandstone to conglomerate indicates either a fall in relative sea level or increase of sediment supply into the system. Incised valleys that back-fill with sandstones may also occur in this area because distal deposits are observed to the north east. X-ray fluorescence (XRF) measurements of the core were acquired in increments of 0.5 ft (0.15 m), except for missing core. The XRF chemical analysis is 16

27 suitable for bulk analysis of major and minor elements in various geologic materials. A handheld Niton XRF analyzer was used to measure the relative quantities of elements for the cored interval. Measured standards analyzed every 20 samples ensured consistency in XRF readings. Important light elements, such as magnesium, aluminum, potassium, silica, calcium, and iron helped further characterize the lithologies in the Marmaton Group. Intervals with high silicon, potassium, and calcium abundances can indicate zones better stimulated by hydraulic fracturing techniques. The Al and K concentration curves mimic each other, suggesting a strong connection between these two elements (Figure 9). Potassium feldspar abundance could be the cause for this correlation; however, the Al concentrations are approximately double that of the K. Ca concentrations are also an order of magnitude higher than K. Although plagioclase has two Al per Ca, this does not account for the sizable difference in elemental abundance. Other volcanic minerals, authigenic clays, and diagenetic minerals (such as chlorite) account for the difference in quantity of elements. Fe and Mg concentrations also follow similar trends. In many minerals, the Fe2+ and Mg2+ ions substitute for each other in crystal structure, thus their abundances would mirror each other. Higher amounts of Fe that do not correspond to an increase in Mg are controlled by iron staining on the surface of the core slab, rather than elemental substitution into minerals. Si concentration are significantly larger than the other light elements measured, especially in sandstone and conglomerate lithologies. Silicate minerals are common in volcanic rocks, which are major constituents in the Granite Wash. Si concentrations spikes occur at the boundaries of sandstone and conglomerate, suggesting a more brittle zone that would be better for hydraulic fracturing. Low Si 17

28 GR Ft Al Ca K Fe Mg Si Figure 9: Concentrations of the six major light elements (Al, Ca, K, Fe, Mg, and Si), vary throughout the cored interval and through the different lithologies. General trends with different elements occur stratigraphically. Around 13,694 ft [4,174 m], at the base of the conglomerate section, a spike is present in all element concentrations. Undulations in abundance occur for each element; however, another noticeable surge in light-element concentration occurs at the beginning of the next conglomerate occurrence, 13,688 ft [4,172 m]. The quantity of light elements then decreases and remains constant, with an exception of a drop of Ca and Mg around 13,691 ft [4,170 m]. At 13,676 ft [4,168 m], all elements increase in concentration except for Fe, which decreases. Low Si values coincide with the start of the shale lithology. All of the measured light elements, except for Si, have low concentrations directly before the shale break. Immediately above the shale, there is a positive shift in elemental concentrations. After the uniform increase of elements, the concentrations oscillate between high and low values, until the concentrations drop towards the top of the cored interval. Transition between sandstone and conglomerate lithologies is often marked by a decrease followed by a sharp increase in light element concentration. 18

29 abundance in the shale may indicate a shallow water level during deposition. Towards the top of the shale, a spike in Ca could signify an increase in calcium carbonate shallow marine life such as corals and/or coccolithophores. 19

30 Electrofacies: Log-based Lithology Estimation Prior to log-based lithology estimation, the wire-line logs were normalized to reduce the differences introduced by various service companies methods of gathering and processing well logs. The GR, DPHI, NPHI, and RHOB well logs were normalized in Geolog (Paradigm). Normalized well logs provided by Devon and Chesapeake were used as the standard to normalize the other well logs using the quantile method (5,50,95). The quantile method is a statistical approach that forces log curves in a given data set to share the same mean value (50th percentile) and scaling for the other given percentiles (5th and 95th) (Shier, 2004). This method takes the assumption that each well contains the same distribution of rock types as the standard log, thus it was employed over smaller areas to minimize errors. The Mayfield 1-34 core description was used to constrain the electrofacies classification (lithology estimation). Because the sample size was small, 82 points over 40 ft (12.2 m), and the Marmaton Group is over 1,000 ft (305 m) thick in some parts, the classification groups were simply conglomerate, sandstone, and shale. More lithology classes would be difficult to estimate. Data from 60 wells from the modeling area (Figure 2), or the immediate area around Mayfield 1-34, were used for the lithology estimation with a k-means cluster analysis. The k-means clustering method (MacQueen, 2006) groups the data from the comparison of rock types from core descriptions and well-log signatures. The algorithm randomly places a user-determined amount of seed points in N-dimensional data space and calculates the means between each data point and the seed point. Movement of the seed points occurs until the distance between data and seed points is minimized (MacQueen, 1967). 20

31 The k-means cluster algorithm classified the data into three different groups. The three assemblages approximately estimate the lithology assessment of the cored interval of Mayfield 1-34 (Table 1). The available well logs to use as parameters for the k-means clustering were: GR, NPHI, DPHI, RHOB, and ILD. Various combinations of the aforementioned well logs (Table 1 and Figure 10) were tested in the k-means cluster algorithm to minimize error from the core description. The combination of NPHI, GR, and ILD produced the best match to the core description with a correlation of 63% (Figure 11). Although the correlation between the electro-lithologies and core-based lithologies is strong, the algorithm had trouble deciphering between very coarse sandstone and granule conglomerate. These two lithologies not only resemble each other visually but also in well-log signatures. The very coarse sandstone and granule-sized conglomerate occur at zones of lithologic transitions within the core, making it more difficult to decipher between the two (Figure 11). Another major issue with this electrofacies scheme is that it characterizes lithologies in zones above Zone E poorly. Well-log signatures change stratigraphically in the Marmaton group; for example, GR readings are overall lower in the top intervals compared to the bottom intervals (Figure 7), which causes misclassification of conglomerates and sandstones compared to Mayfield 1-2 and Sage 1-34H physical core descriptions (Table 2). The Mayfield 1-2 well was described as being entirely coarse-very coarse sandstone with occasional intervals of silty finegrained sandstone deposits; the algorithm classified the entire interval as conglomerate. The Sage 1-34H core was not described continuously but rather at defined points. Porosity and permeability measurements were taken at distinct points as well. However, 21

32 NPHI (V/V) Cluster GR (API) ILD (ohm.m) # Data Points Mean Std Dev. Mean Std Dev. Mean Std Dev. Conglomerate Sandstone Shale Table 1: Data distributions used to define the three main lithologies: conglomerate, sandstone, and shale. In the k-means algorithm, three log curves (NPHI, GR, and ILD) were examined in 3-dimesional space. Random seed points, or cluster centers, were chosen for three groups (the three lithologies) and the first round of clusters were generated. Subsequently, the algorithm reseeds the center points to minimize the distance from the data points to the seed points of the clusters. This process is repeated until the shortest distance from the data points to the center was achieved. In total, 83 data points were used. 22

33 A B ILD (ohm.m) ILD (ohm.m) C NPHI (v/v) GR (API) NPHI (v/v) Key Conglomerate data point Conglomerate seed point Sandstone data point Sandstone seed point Shale data point Shale seed point GR (API) Figure 10: Cross plots of A) NPHI vs. ILD, B) GR vs. ILD, and C) GR vs. NPHI. Using the 40 ft (12.2 m) of core from the Mayfield 1-34 well, k-mean algorithm cluster analysis was performed to estaphblished electrofacies. Three logs were used: GR, ILD, and NPHI. The data points have been classified into three clusters: conglomerate, sandstone, and shale. A) The cross plot of NPHI vs. ILD show a concentrated grouping of points for the conglomerate with increasing cluster space from sandstone to shale. B) The cross plot of GR vs. ILD values shows dispersed clusters for the three lithologies. The most variance of data point values occurs with sandstone and shale. C) Cross plot of NPHI and GR values. The conglomerate and sandstone data points form concentrated clusters around their seed points, while the shale data points have a wide spread of NPHI values ( ) and GR values ( ). 23

34 GR Ft Core Lithologies Electrofacies Lithologies Cutoff Lithologies Figure 11: GR, core lithologies, electrofacies lithologies, and cutoff lithologies for the Mayfield 1-34 well. The electrofacies lithologies were calculated using the k-mean clustering algorithm. It does not capture the thinner beds of lithologies such as the thin conglomerate interval around 13,667 ft [4,166 m]. The shale break is predicted to be larger with the electrofacies than what was physically seen in core, most likely due to missing shale segments in retrieval. There is also a predicted sandstone package immediately following the shale lithology that is not seen in core. It is described as a conglomerate containing mostly pebble sized clasts, which the algorithm classified as a very coarse sandstone. That mismatch of predicted electrofacies lithologies and actual lithologies is seen to the bottom in the cored interval. The cutoff lithologies do not show the thin conglomerate around 3,667 ft [4,166 m] as well. It does decipher between the conglomerate and sandstones lithologies similarly to the physical description of the core. Overall, the cutoff predicted lithologies match the core lithologies better than the electrofacies predicted lithologies. 24

35 Well name Depth (ft) Zone Core Lithology Predicted Lithology Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Mayfield 1-2 Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Sage 1-34 H Very coarse sandstone Coarse sandstone Very coarse sandstone Coarse sandstone Very coarse sandstone Very coarse sandstone Very coarse sandstone Very coarse sandstone Very coarse sandstone Very coarse sandstone Silty shale Coarse sandstone Very coarse sandstone Coarse sandstone Very coarse sandstone Very coarse sandstone Very coarse sandstone Very fine sandstone Silty shale Very fine sandstone Very fine sandstone Very fine sandstone Fine sandstone Very fine sandstone Very fine sandstone Fine sandstone Fine sandstone Fine sandstone Fine sandstone Fine sandstone Fine sandstone Very fine sandstone Very fine sandstone Very fine sandstone Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Conglomerate Sandstone Sandstone Conglomerate Sandstone Conglomerate Conglomerate Conglomerate Conglomerate Sandstone Sandstone Sandstone Sandstone Sandstone Sandstone Sandstone Shale E E E E E E E E E E E E E E E E E Wash Wash A B B B B B B B B B B B C C C Table 2: Point comparisons of actual lithologies versus predicted lithologies. For the Mayfield 1-2 well, the algorithm assigned every point to be conglomerate, not recognizing finer grained intervals. The Sage 1-34 core has predominately very fine grained materials that the algorithm does not always classify correctly. There are predicted conglomerates when there are actually very fine grained sandstones. 25

36 comparing the predicted lithologies to the observed lithologies at these points shows a mismatch and an overall correlation of 53%. Again, the algorithm over estimates the amount of conglomerate compared to the observed amount. To use this algorithm effectively, it should be applied to the Marmaton F interval only. In order to properly classify the changing characteristics of lithologies in each zone, a cored section containing all three lithologies and its corresponding well logs would be needed to constrain the algorithm. Lithology logs were calculated using cut-off values (Table 3) to compare to the electrofacies logs. Shale value cutoffs were determined by examining the GR log values per zone and locating shale base-lines. Then looking at the GR histograms, finding the lowest GR value that corresponds to the shale group. Separating sandstone and conglomerate lithologies proved to be more difficult. It involved utilizing core descriptions from Duggins (2013) and core data from the Sage 1-34 H well, and assumed that the lower and upper zones had similar characteristics. The log-cutoff technique, a less mathematically precise method, matched lithologies better for the whole Marmaton Group compared to the k-means clustering technique. A major part of the success was due to the zones being treated individually rather than a homogenous interval. Compared to the Mayfield 1-34 core description, the lithology log accurately predicts (Figure 11) the lithology with a correlation of 74%. The Mayfield 1-2 and Sage 1-34 had 100% and 76% correlations respectively. 26

37 Zone Shale Sandstone Conglomerate Wash GR 105 RHOB > 2.52 RHOB 2.52 A GR 105 RHOB > 2.57 RHOB 2.57 B GR 105 RHOB > 2.55 RHOB 2.55 C GR 110 RHOB > 2.52 RHOB 2.52 D GR 120 RHOB > 2.52 RHOB 2.52 E GR 125 RHOB > 2.52 RHOB 2.52 F GR 135 RHOB > 2.56 RHOB 2.56 Table 3: Well-logs used for the lithology logs and the value cut-offs for each lithology. 27

38 Structural and Stratigraphic Framework Stratigraphic correlations of 430 wells (353 wells with digital logs and 77 wells with raster logs) were used to construct the structural and stratigraphic framework of the Marmaton Group. Wire-line log signatures for GR, ILD, NPHI, DPHI, and RHOB signatures guided top picks, as well as for 76 wells with formation top picks for the Marmaton Group. The Mayfield 1-34 well log and formation tops serves as a type log for the general trends and formation tops interpreted throughout the area of interest (Figure 7). The tops include: Marmaton Top, Marmaton A, Marmaton B, Marmaton C, Marmaton D, Marmaton E, Marmaton F, and the Upper Skinner Shale (or Marmaton Base). These tops define the seven zones or formations: Top, A, B, C, D, E, and F. Regional cross sections (perpendicular and parallel to the Mountain View Fault), which contains the type log, were correlated first to identify basin-wide flooding surfaces and trends (Figures 12 and 13). The basin-wide flooding surfaces have a combination of high GR readings (can be over 200 API), high NPHI readings (up to 0.2 v/v, corresponds to higher organic content), and increases in ILD (up to 200 ohm.m). Their wire-line log characteristics show a greater change in the aforementioned log values than other, minor shales within the Marmaton Group that are usually localized. Then smaller, local area cross sections, also perpendicular and parallel to the Mountain View Fault, were used to make short-range formation top interpretation. These smaller scale cross sectional areas were correlated from west to east (Appendix F). Regional cross sections spanning NW-SE and N-S show trends of porosity, GR, and thickness that the Marmaton Group, as well as individual zones follow (Figures 12 and 13). Most of the interpreted faults have strike azimuths of NE-SW which indicates 28

39 Depth (ft) A TX State Line OK A Marmaton Wash Marmaton A Marmaton B Marmaton C Marmaton D Marmaton E Marmaton F Figure 12: This structural cross section (refer to Figure 2 for location) has equally spaced wells that are not to scale. Structure elevation decreases to the southeast and zones have much higher GR readings. The red lines denote possible regional reverse faults with displacements of 600 ft (183 m). GR

40 Depth (ft) B B Marmaton Wash Marmaton A Marmaton B Marmaton C Marmaton D Marmaton E Figure 13: This structural cross section of the Marmaton Group (refer to Figure 2 for location) has equally spaced wells that are not to scale. Generally, the structural elevation of the formation tops increase going south and zones Wash-F increase in thickness. The red lines denote possible reverse faults shows possible reverse faulting as well, with displacements between ft ( m). Marmaton F GR

41 that they could be reactivated faults from the Oklahoma aulacogen (Figure 14). Thus, the greatest apparent offsets are seen in the NW-SE directions and can have up to 600 ft (183 m) of separation (Figure 12). High GR readings are prevalent in the southern portion of the study in Beckham County (Figures 12 and 13) that correspond to greater amounts of shale and/or potassium minerals as seen in the previous XRF data. Zone thickness is greatest adjacent to the Mountain View Fault with an exception that is bound by faults (Figure 14). The thinnest deposits occur in the northeastern area of the study area and due to lack of well control in the area it is unknown if it is fault related. Using the interpreted horizons (tops) from core and well data, a 3-D stratigraphic and structural framework (3-D reservoir model grid) was constructed for the Marmaton Group (Figure 15) of dimensions 11.9 mi x 8.33 mi x 0.49 mi (19.2 km x 13.4 km x 0.79 km). The 3-D grid has individual cells that are 250 ft by 250 ft (76.2 m by 76.2 m) aerially and approximately 1.5 ft (0.46 m) vertically, for a total of 47.2x106 cells. The layering (stratal geometry) within each of the seven zones was created using proportional layers. The stratigraphy follows the structure of the basal Marmaton Group, suggesting that large scale tectonics influenced the system (reverse faulting and folding) throughout the Desmoinesian (Figure 16). Overall, the bottom zones (C-F) are thicker than the top zones (Wash-B) but there are intra-zonal thickness changes. As seen in Figure 16, there are possible erosional events and large-scale growth units. This indicates a non-uniform rate of uplift and subsidence during the time of deposition. Lithology logs based on the well-log cutoffs were upscaled to the model grid. Because this is a discrete property, the upscaled cell lithology was determined by the highest occurring lithology within a cell. A visual representation lithology proportions by layer 31

42 A Structure map for the top of the Marmaton Group N Wheeler Roger Mills Depth (ft) Beckham mi km B Isopach map for the Marmaton Group N Thickness (ft) Wheeler Roger Mills Beckham 0 0 mi km Figure 14: A) Structure map for the top of the Marmaton Group and B) isopach map for the entire Marmaton Group. A) The structure map shows a trend of increasing structural elevation to the northwest. The deepest elevations occur on the county border between Beckham and Roger Mills counties which coincides with the axis of the basin. The red dashed lines are possible faults within the study area. B) The isopach map for the Marmaton Group shows patterns of thick and thin deposits that line up with the interpreted faults. The thin syndepositional deposits correspond to the up-thrown side of the reverse faults, while thick syndepositional deposits correspond to the down-thrown side. Thicker sediment accumulations occur along the southern boundary of the study area, which coincides with the Amarillo-Wichita uplift. 32

43 A Zones Marmaton Wash Marmaton A Marmaton B Marmaton C Marmaton D Marmaton E Marmaton F V.E. 1x B Zones Marmaton Wash Marmaton A Marmaton B Marmaton C Marmaton D Marmaton E Marmaton F V.E. 5x Figure 15: A) Wells with KB surface and unexaggerated model. B) Exaggerated model. A) 56 vertical and 4 horizontal wells were used as guides for the modeling. B) The exaggeration model grid shows in greater detail changes in zone thicknesses. The model grid dimensions are approximately 11.9 mi x 8.33 mi x 0.49 mi (19.2 km x 13.4 km x 0.79 km) for a total of 47.2x10 6 cells. 33

44 A Marmaton Bottom Depth (ft) Y-axis X-axis Zones Marmaton Wash Marmaton A Marmaton B Marmaton C Marmaton D Marmaton E Marmaton F Z-axis X-axis Y-axis V.E. 7.5x B B A 0 mi 1 V.E. 7.5x 0 km 1.6 Figure 16: A) An i slice of the zone model laid on top of the structure map for the bottom of the Marmaton and B) an eastern-view of the i slice. A)The lower zones (C-F) are the thickest in the Marmaton Group, while the upper zones (Wash-B) are significantly thinner. B) Zones follow the topography for the base of the Marmaton Group suggesting that large scale structural events influence all zones (red dashed fault). Box A represents a possible erosion event in which the Marmaton F was eroded in an area and the Marmaton E sediment filled the accomodation space. Box B shows a possible growth unit in the Marmaton C. The Marmaton D does not show thickness changes suggesting a cease of fault movement but it starts again in the Marmaton C. 34

45 (stratigraphically) is shown as the vertical lithology proportion curve (Figure 17). The vertical lithology proportion curve for the Marmaton Group can be interpreted as depositional cycles; pulses of conglomerate, then a fining up sequence (increase in sandstone), then a pronounced shale deposit. There are thicker coarse deposits (conglomerate and sandstones) in the stratigraphically lower Marmaton zones (E and F), while the upper Marmaton zones have higher percentages of shale. For zone F, there were two conglomeritic intervals with no distinct shale interval separating them. This could mean that the shale was either not deposited in this time frame or, more likely, it was eroded by the following substantial influx of conglomerate. Moving up stratigraphically, the conglomeritic packages decrease in thickness and sandstone becomes the predominate lithology. Consequently, the marine shales and flooding surfaces occur more frequently (Figure 17). Zone B has a bell-shaped conglomerate proportion potentially indicative of a complete cycle of coarsening upward and fining upward. The top two zones exhibit uniform fractions of each lithology. Interpretation of these cyclical deposits are marine regressions (coarse deposits) and transgressions (fine deposits). Although more evidence (geochemical and structural) is needed to definitively determine the controls on the regressive and transgressive sequences, uplift, weathering, and erosion providing sediment influx and movement along reverse faults are most likely the major high-order controls on the system. The sediment shedding and a slowing of fault slip could fill the basin and diminish accommodation space, thus promoting marine regression. Basin subsidence could create more accommodation space that the sediment influx cannot fill. This would 35

46 Lithology Proportion (fraction) Wash A B MFS C MFS Layer D MFS 700 E MFS F Figure 17: Vertical proportion curve of the three different modeled lithologies by layer and zone. Orange is conglomerate, yellow is sandstone, and grey is shale. MFS stands for marine flooding surface. There are pulses of conglomerate followed by shale caps, indicating a cyclic deposition controlled by relative rise and fall of sea level. 36

47 cause higher relative sea level locally and this quiescent environment would produce shale deposits. 37

48 Spatial Distribution of Lithology The 3-D stratigraphic and structural framework (3-D model grid of the Marmaton Group stratigraphic zone) was used as a first-order constraint to map the spatial distribution of lithology, effective porosity, and water saturation in the study area. Sequential-indicator simulation (SIS) was used to generate a constrained lithology model using 1) upscaled well logs, 2) histogram of lithology percentages, 3) vertical and horizontal variograms of each lithology (by zone), and 4) the vertical lithology proportion curve. Based on the upscaled wells, the percentages of each lithology that are honored in the lithology model are 23 % conglomerate, 59% sandstone, and 18% shale. Vertical and horizontal variograms were constructed by zone to analyze and identify spatial variability for each lithology (Appendix G). The major and minor anisotropy and the azimuth of the greatest anisotropy were measured using horizontal variogram maps. In addition at looking at the raw data to decide vertical and lateral ranges, information about alluvial fan and fan delta deposits were used to critical evaluate reasonable ranges for each lithology. Conglomerate deposits in these environments are relatively vertically continuous but are not as laterally extensive. Sandstone deposits exhibit vertically continuous deposits but are both laterally continuous and discontinuous patterns. Shales are discontinuous vertically but have extensive lateral deposits. This knowledge guided the values inferred from both the vertical and horizontal variograms. The vertical ranges for conglomerates, sandstones, and shales are as follows: ft ( m), ft ( m), and m) respectively. The major and minor horizontal ranges are considerably 38

49 larger and their ranges for the different lithologies as follows: for conglomerates 5,2007,700 ft (1,585-2,347 m) major and 2,600-6,000 ft (792-1,829 m) minor, for sandstones 3,600-11,500 ft (1, m) major and 1,800-7,000 ft (549-2,134 m) minor, and for shales 7,000-12,000 ft (2,134-3,658 m) and 3,000-7,000 ft (914-2,134 m) minor. Lithology percentages by zone (Appendix H) show that zones Wash, D, E, and F are dominated by conglomerates and sandstones and that zone B has approximately 35% shale. This corresponds to first estimates of the thicker zones having higher percentages of conglomerates and sandstones. This lithology model (Figure18) also shows how the lithology deposits are distributed. Layer 880 (Marmaton F) has a large proportion of sandstones and conglomerates. A possible interpretation for that is sandstone channels with a north by northeast depositional trend are well developed but the conglomerate has a more amorphous, continuous deposit shape that is trending in the northeast direction (Figure 18). A marine shale below a flooding surface, layer 587 (Marmaton E), shows the laterally continuous shale deposits that are expected for fan delta deposits (Figure 18). Although there is significantly less sandstone than the previous interval, a north by northeast possible channel depositional trend exists. Channels trending to the northeast make sense if they are being fed by the Amarillo-Wichita uplift and sediment would be transported in that direction. The combined conglomerate and sandstone lithology percent maps were constructed to observe patterns for sediment shed off of the Amarillo-Wichita uplift (Figure 19). Zones with high conglomerate and sandstone content show channelized sediment orientations that are approximately perpendicular to the Mountain View Fault. This agrees with the type of dendritic channelized sandstone and conglomeritic deposits 39

50 A Lithology -Conglomerate -Sandstone -Shale V.E. 5x B N C N 0 mi 2 0 mi 2 0 km km 3.22 Figure 18: Lithology model two stratigraphic lithology representations. A) Shale deposits appear as more laterally continuous deposits throughout the layers, while the sandstone is more channelized as expected with the associated fan delta deposits. Conglomerate deposit patterns depend on where in the cycle the system is. During high conglomerate deposition, it blankets the layers, where as in periods of decreased sediment input, channelized flows can be seen. B) A layer from the Marmaton F zone where conglomerate and sandstone deposits are dominant (k slice 880). The sandstone deposits are interpreted to form channels in the north by northeast direction (arrows) while the conglomerate deposits are more laterally connected trending in the northeast direction. C) Shale is the dominate lithology in this layer (k slice 587) and appears very continuous with no apparent depositional trend azimuth. The sandstone bodies are too disconnected to decipher an azimuth. 40

51 A Conglomerate and Sandstone % 100 Zone F N mi 2 B 0 km 3.22 Conglomerate and Sandstone % Zone B N 0 mi 2 0 km 3.22 Figure 19: Conglomerate and sandstone percent maps. A) There is very little shale lithology throughout Zone F, however, a faint trend of north by northwest of conglomerate and sandstone deposits is seen. Although the trend is not perpendicular to the Wichita-Amarillo uplift, it is still within the realm of possibility. Current and wind directions could feasibly shift sediment transportation this much. B) Zone B contains a large portion of shale. The conglomerate and sandstone depositional trend has an east-west linear pattern. 41

52 found in alluvial fans and fan deltas. High shale-bearing zones, such as zone B, show sinuous thick deposits of sandstone and conglomeritic deposits as well. In every nonshale lithology percent map, the area juxtaposed to the Amarillo-Wichita uplift contains between 80 and 100% sandstone and conglomerate. This area is close to the source so natural sorting of sediments would favor coarser-grained material in this region and this is the axis of the Anadarko Basin. Also continued uplift of the Mountain View Fault and subsidence of the basin throughout the Desmoinesian created a large amount of accommodation space for sandstones and conglomerates. In order to assess reservoir quality for the Marmaton Group, Vshale, ϕt (total porosity), ϕe (effective porosity), and Sw (water saturation) were calculated using the GR, NPHI, DPHI, and ILD wire-line logs for each of the seven zones (Appendix I). Vshale was calculated using the GR method and a root mean square of NPHI and DPHI resulted in ϕt. Both Vshale and ϕt were used to calculate ϕe, which in turn was used in the equation for Sw. Sequential-gaussian simulation (SGS) models for ϕe and Sw were constrained by 3-D lithology model. The first step to constructing these models was to upscale their logs. Vertical and horizontal variogram ranges for created for conglomerate or sandstone (Appendix G). The ranges had to be shorter than the lithology variance ranges because there is internal heterogeneity for each lithology. Conglomerate vertical ranges were from 14.8 to 23.8 ft (4.51 to 7.25 m) for ϕe and 11 to 18.6 ft (3.35 to 5.67 m) for Sw. Sandstone vertical ranges spanned from 12.5 to 21.7 ft (3.81 to 6.61 m) for ϕe and ft (4.39 to 5.43 m) for Sw. Horizontal major and minor ranges were kept constant regardless of property of lithology by zone and they ranged from 2,600 to 3,850 ft (792.5 to 1,173 m) for the major direction and from 1,300 42

53 to 3,000 ft (396.2 to m) for the minor direction. Shale cells were set to ϕe=0 and Sw=100 using the assumption that all the shale porosity is occupied by clay-bound water and/or is not reservoir. Normal-score transformations were applied to the petrophysical properties by zone and were used as a constraint for the SGS modeling. The resulting models for ϕe and Sw show the spatial trends (Figure 20). Arithmetic means for ϕe and Sw were calculated for each lithology (conglomerate and sandstone) by zone (Appendix H). Conglomerates have a higher average ϕe and lower average Sw for the Marmaton Group (and individual zones) than sandstones. The highest ϕe and lowest Sw values for conglomerates occur in zone E. Water saturation values are close for both sandstones and conglomerates except in zones D-F, which there is a low abundance of shale. Four intervals with maximum transgressive (high shale content) and regressive (high sandstone and/or conglomerate content) units were compared by lithology, ϕe, and Sw spatial distributions (Figure 21). Overall trends for ϕe and Sw values corresponding to the transgressive-regressive cycles are higher porosity and lower water saturation in the regressive stage. Less shale and high variability of deposits created the speckled pattern seen in both petrophysical properties. Transgressive surfaces contain more shale. This leads to lower ϕe and higher Sw values even in the non-shale lithologies. Presumably the sandstones and conglomerates mixed with mud during these periods thus decreasing reservoir quality. Inspecting these surfaces individually provides more information about stratigraphic variability of the petrophysical properties. The bottom interval (Marmaton F) is a major regressive package with a large amounts of conglomerates. Visually, the sandstone patches are a dark blue to purple in the ϕe model indicating low porosity. The 43

54 A ϕ e V.E. 5x B Sw % V.E. 5x Figure 20: Effective porosity and water saturation models. A) The entire effective porosity model. Shale lithologies were given a value of zero, thus the effective porosities for conglomerates and sandstones are highlighted throughout. B) The water saturation model also assigned a universal value to shales: 100%. Again, this allows for the focus of analyses to be on the conglomerates and sandstones. 44

55 Lithology Proportion (fraction) Wash A B C 100 A 200 B 300 C D Layer E F Lithology -Conglomerate -Sandstone -Shale 0 0 mi 5 km 8.05 ϕ e mi 5 km 8.05 Sw % mi 5 km 8.05 Figure 21: Comparisons between A) lithology, B) effective porosity, and C) water saturation models. Layers 880 and 379 show times of maximum regression, thus high percentages of sandstone and conglomerate. In these slices there are higher effective porosities and lower water saturations. Layers 587 and 147 show points of maximum transgression, thus shales dominate. This drives down the average effective porosity closer to 0 and water saturation up to 100%. 45

56 Sw values for this interval mostly lie between 0-20% with occasional areas of higher values. The next stratigraphic surface represents a major transgressive event in which expansive sheets of shale and minimal conglomeratic material were deposited. Sandstone contacts with shale show lower effective porosity and higher water saturation. However, the inner area of these sandstone localities have higher ϕe and lower Sw values. This could possibly mean that the peripheral sands mixed with mud for a longer duration or the shale could be acting as a baffle. The third stratigraphic interval (Marmaton D) captures a marine regression. There are some conglomerate packages emanating from the edge of the model area that appear to have warmer color swarms (higher ϕe values) but the massive sandstone also has a few higher porosity trends. The dark blue (low porosity) cells meanders through the sandstone sheets create a distinct, connected path throughout the model area. Water saturation does not have any unique patterns associated with conglomerates or sandstones. Like the other regressive zone, this interval has between 0-20% water saturation, with higher values near shale contacts. The last stratigraphic interval (Marmaton B) is a transgressive event. Little conglomerate deposits occur throughout the area but the sandstone and shale have a west-east depositional orientation. 46

57 Conclusions The Desmoinesian Granite Wash, specifically the Marmaton Group, is a hydrocarbon-bearing interval within the Anadarko Basin of Oklahoma and Texas that is composed of clastic and carbonate sediments derived primarily from the AmarilloWichita Uplift. The Marmaton Group, located in Beckham County, Oklahoma and Wheeler County, Texas, includes a series of vertically stacked conglomerates and tightgas sandstones and shales that exhibit a complex stratigraphic architecture, highly variable lithologies, and correspondingly heterogeneous reservoir properties. The stratigraphic and reservoir characteristics of the Marmaton Group, are established based on cores, x-ray fluorescence (XRF) measurements, and well-log signatures. The Marmaton Group in the southern Anadarko Basin contains interbedded arkosic sandstones and conglomerates that thin laterally into shales to the north (basinward). At least four regional, correlatable flooding surfaces (and associated organic-rich shales) subdivide the Marmaton Group and are thought to be self-sourcing in this liquids-rich interval. Proximal to the Amarillo-Wichita uplift, the Marmaton Group is highly lithologically heterogeneous. XRF analyses of cored intervals show that elemental concentrations vary stratigraphically in conjunction with lithology. XRF measurements show that concentrations of potassium and aluminum have the same increasing and decreasing trends suggesting that they may be associated (in minerals such as potassium feldspar). Characteristic well-log signatures correspond to different intervals and can be correlated laterally through the study area. Cluster analysis implemented on well-log data resulted in a 63% correlation to the Mayfield 1-34 core description but achieved 47

58 low correlations for the Mayfield 1-2 (0%) and Sage 1-34H (53%). Well-log cutoffs performed on well-log data have a 74% correspondence rate to Mayfield 1-34 core description. Overall the well-log cutoff lithologies provides an approximation of lithologies in non-cored: 62% sandstone, 23% conglomerate, and 15% shale. A compiled lithology model of the Marmaton Group displays spatial patterns by zone constrained to the vertical lithology proportion trend, vertical variograms, horizontal variograms, and lithology percentages. Sandstone and conglomerate deposits appear to have a dendritic channel trend perpendicular/sub-perpendicular to the Amarillo-Wichita uplift. Shale deposits display a more laterally continuous, vertically discontinuous trend with no discernable depositional azimuth trend. Using the lithological trends as an input, effective porosity and water saturation show that conglomerates on average have a higher effective porosity (by 1%) lower water saturation (by 1%) throughout the Marmaton Group. 48

59 References Blakey, R., 2013, North American Key Time-layer Paleogeographic Maps, Accessed 08/18, Brown Jr, L. F., 1979, Deltaic sandstone facies of the Mid-Continent, in Hyne, N. J., ed., Pennsylvanian Sandstones of the Mid-Continent, Tulsa, OK, Tulsa Geological Society, vol. 1, p Campbell, J. A., C. J. Mankin, A. B. Schwarzkopf, and J. J. Raymer, 1988, Habitat of petroleum in Permian rocks of the midcontinent region; in, Permian Rocks of the Midcontinent, W. A. Morgan and J. A. Babcock, eds.: Midcontinent Society of Economic Paleontologists and Mineralogists, Special Publication No. 1, p Doveton, J. H., 1994, Multivariate Pattern Recognition and Classification Methods: Chapter 4, in Geologic Log Analysis Using Computer Methods, Tulsa, Oklahoma, AAPG, vol. 2, p Duggins, W. T., 2013, Facies Architecture and Sequence Stratigraphy of Part of the Desmoinesian Granite Wash, Texas Panhandle and Western Oklahoma: Master s Thesis, University of Tulsa, Tulsa, Oklahoma, 166 p. Dutton, S. P., 1984, Fan-Delta Granite Wash of the Texas Panhandle, Oklahoma City Geological Society, vol. Short Course Notes, p Eddleman, M. W., 1961, Tectonics & Geologic History of the Texas & Oklahoma Panhandles, in Wagner, C. R., ed., Oil and gas fields of the Texas and Oklahoma panhandles, Amarillo, TX, Panhandle (Texas) Geological Society, p Edwards, A. R., 1959, Facies changes in Pennsylvanian rocks along north flank of Wichita Mountains, in Ardmore Geol. Soc., Petroleum geology of southern Oklahoma-a symposium, Tulsa, Oklahoma, AAPG, vol. 2, p Esbensen, K. H. and A. G. Journel, 1992, Multivaiate Data Analysis: Part 6.Geological Methods, in Morton-Thompson, D. and A. M. Woods, eds., Methods in Exploration: Development Geology Reference Manual, Tulsa, Oklahoma, AAPG, vol. 10, p Graham, S. A., W. R. Dickinson, and R. V. Ingersoll, 1975, Himalayan-Bengal Model for Flysch Dispersal in the Appalachian-Ouachita System, Geological Society of America Bulletin, vol. 86, no. 3, p Ham, W. E., R. E. Denison, and C. A. Merritt, 1965, Basement Rocks and Structural Evolution of Southern Oklahoma--A Summary, AAPG Bulletin, vol. 49, no. 7, p

60 Hoek, E. and E. Brown, 1997, Practical estimates of rock mass strength, International Journal of Rock Mechanics and Mining Sciences, vol. 34, no. 8, p Jain, A. K., 2010, Data clustering: 50 years beyond K-means, Pattern Recognition Letters, vol. 31, no. 8, p Johnson, K. S., 1989, Geologic evolution of the Anadarko Basin, in Johnson, K. S., ed., Anadarko Basin symposium, Atlanta, GA, Oklahoma Geological Survey Circular, vol. 90, p Johnson, K. S. and K. V. Luza, 2008, Earth sciences and mineral resources of Oklahoma, Educational Publication 9, Oklahoma Geological Survey, 22 p. Kaźmierczuk, M. and J. Jarzyna, 2006, Improvement of lithology and saturation determined from well logging using statistical methods, Acta Geophysica, vol. 54, no. 4, p Keller, G., E. Lidiak, W. Hinze, and L. Braile, 1983, The role of rifting in the tectonic development of the midcontinent, USA, Tectonophysics, vol. 94, no. 1, p Kluth, C. F. and P.J. Conet, 1981, Plate tectonics of the Ancestral Rocky Mountains, Geology, vol. 9, no. 1, p Laubach, S. E., J. E. Olson, and M. R. Gross, 2009, Mechanical and fracture stratigraphy, AAPG Bulletin, vol. 93, no. 11, p LoCricchio, E., 2012, Wash Play Overview, Anadarko Basin: Stratigraphic Framework and Controls on Pennsylvanian Granite Wash Production, Anadarko Basin, Texas and Oklahoma, AAPG Search and Discovery Article, no MacQueen, J., 1967, Some Methods for Classification and Analysis of Multivariate Observations, Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, p McConnell, D. A., 1989, Determination of Offset Across the Northern Margin of the Wichita Uplift, Southwest Oklahoma, Geological Society of America Bulletin, vol. 101, p McGowen, J., 1971, Alluvial Fans and Fan Deltas: Depositional Models for Some Terrigenous Clastic Wedges, AAPG Bulletin, vol. 55, no. 1, p Merkel, R. H., 1979, Responses of the Three Porosity Logs in More Complex Lithologies, in Anonymous Well Log Formation Evaluation, Tulsa, OK, American Association of Petroleum Geologist, p

61 Mitchell, J., 2011, Horizontal Drilling of Deep Granite Wash Reservoirs, Anadarko Basin, Oklahoma and Texas, Shale Shaker, vol. 62, p Moore, G. E., 1979, Pennsylvanian paleogeography of the southern MidContinent, in Hyne, N. J., ed., Pennsylvanian Sandstones of the Mid-Continent, Tulsa, OK, Tulsa Geological Society, vol. 91, p Northcutt, R. A. and J. A. Campbell., 1995, Geologic provinces of Oklahoma: Oklahoma Geological Survey Open-File Report 5-95, 1 sheet, scale 1:750,000, 6-page explanation and bibliography. Ratcliffe, K., J. Martin, T. Pearce, A. Hughes, D. Lawton, D. S. Wray, and F. Bessa, 2006, A regional chemostratigraphically-defined correlation framework for the late Triassic TAG-I Formation in Blocks 402 and 405a, Algeria, Petroleum Geoscience, vol. 12, no. 1, p Reading, H. G. and M. Richards, 1994, Turbidite systems in deep-water basin margins classified by grain size and feeder system, AAPG Bulletin, vol. 78, no. 5, p Sahl, H. L., 1970, Mobeetie field, Wheeler County, Texas, Shale Shaker, vol. 20, p Serra, O., and H. T. Abbott, 1980, The contribution of logging data to sedimentology and stratigraphy: SPE 9270, 55th Annual Fall Technical Conference and Exhibition, Dallas, Texas, 19 p. Shier, D.E., 2004, Well Log Normalization: Methods and Guidelines, Petrophysics, vol. 45. No. 3, p Wickham, J., 1978, The Southern Oklahoma Aulacogen, Structure and Stratigraphy of the Ouachita Mountains and the Arkoma Basin: Field Guide, p Wirth, K. and A. Barth, 2013, X-Ray Fluorescence (XRF), Accessed 04/18,

62 Appendix A: Core Description 52

63 Depth Gravel Sand Mud Sorting Cobble Pebble Granule VC Sand C Sand M Sand F Sand VF Sand Clay (ft) P M G Lithology Color Description X Conglomerate dark grey/charcoal matrix, openish matrix, extremely varing clast size, largest about 20mm, fine grained volcanixs, ss light grey, closed matrix, iron staining, clasts up to 55mm Missing core Missing core X X mixed ss dark grey/dark tan tan tan/dark grey flaser to lenticular very coarse ss, massive flaser going up to lenticular Missing core Missing core X ss light grey, dark grey/ charcoal Fining upward with mud ripples towards top X med grey v coarse ss and dark mud ripples, wavy bedding, some mud clasts X mud stone ss tan/dark grey Wavy bedding about 50/50 sand to mud, fine ss X conglomerate light grey Fining upward 6mm-<1mm Missing core Missing core 53

64 Depth Gravel Sand Mud Sorting Cobble Pebble Granule VC Sand C Sand M Sand F Sand VF Sand Silt Clay (ft) P M G Lithology Color Description X mixed dark grey mud, light grey sand c sand-clay coarser on bottom, mud rip ups extremely prevelant, fining upward in background?, open framwork, extremely distorted bedding (debris flow?) X ss light grey Massive ss with rounded, sphereical grains Missing core Missing core Shale dark grey, charcoal, light black Missing core Missing core Shale dark grey, charcoal, light black Missing core Missing core Shale Dark Grey Tan silt laminations and lenses

65 Depth Gravel Sand Mud Sorting Cobble Pebble Granule VC Sand C Sand M Sand F Sand VF Sand Silt Clay (ft) P M G Lithology Color Description Shale Dark Grey Tan silt laminations and lenses x Conglomerate light tan-tan larger clast >20mm, mostly fining upward sequence Missing core Missing core X Conglomerate light grey Fining upward 30mm-1mm Missing core Missing core X Conglomerate light grey ss Missing core Missing core 55

66 Cobble Pebble Granule VC Sand C Sand M Sand F Sand VF Sand Silt Clay Depth Gravel Sand Mud Sorting (ft) P M G Lithology Color Description X Conglomerate light grey closed matrix, slightly iron stained, same are conglom before missing section "finer" grained sharp contact on top, coarsening upward Missing core Missing core X Conglomerate Light Grey Very poorly sorted and smaller clasts, broken clasts in the matrix up to 50mm clasts, closed matrix light grey rounded -sub angular mixed spherecity, iron stained, some marble? Fine grained volcanics as well

67 Cobble Pebble Granule VC Sand C Sand M Sand F Sand VF Sand Silt Clay Depth Gravel Sand Mud Sorting (ft) P M G Lithology Color Description X Conglomerate Light Grey 50mm clasts, closed matrix light grey rounded -sub angular mixed spherecity, iron stained, some

68 Cobble Pebble Granule VC Sand C Sand M Sand F Sand VF Sand Silt Clay Depth Gravel Sand Mud Sorting (ft) P M G X Lithology ss Color dark grey mud and light tan sand Description Flat ripples, rounded sphereical pebbles entrained Inclined, climbing ripples, fair amount of mud but not 50/50 styolite at bottom. Lenticular bedding X Conglomerate dark grey mud and light tan sand sharp contrast "finer" conglom light gray closed matrix bottom and "coarser" open charcoal matrix on top Cobble + size fine grained volcanic Missing core X Missing core Conglomerate Conglom on wavy bedding (sediment loading), closed matrix mud on top, X ss-clay dark grey clasts c sand - pebble mud and light Wavy bedding about 50/50 sand to Light grey close matric rounded to tan f-vf sand subangular clasts low spherecity. "Finer" grains med sand-vc coarse sand. Mud drapes "finer" grains med sand-vc coarse sand X Conglomerate dark grey mud clasts some light tan silt within Light grey close matric rounded to subangular clasts low spherecity, up to 50mm volcanic clasts dark greycharcoal Missing core Missing core 58

69 Depth Gravel Sand Mud Sorting Cobble Pebble Granule VC Sand C Sand M Sand F Sand VF Sand Silt Clay (ft) P M G Lithology Color Description Missing core Missing core X ss- conglom Light Grey, dark greycharcoal Open matrix rounded mostly low spherecity, scour surface, volcanic frags granodiorite? 5mm-35mm clasts smaller clasts are ss and sedimentary, X X ss Light Grey Mud drapes fining upwards up to fine sand, styolite? Mud clasts Coarsening up, mud clasts on bottom 59

70 Appendix B: XRF Mechanics X-rays emitted from the XRF spectrometer excite atoms of different elements and can remove an electron from an inner electron shell (Wirth and Barth, 2013). The instability created by this disruption of shell configuration causes an electron from an outer shell restores the inner shell moving to the ejected electrons place. This replacement releases energy due to the lower energy needed to hold an electron in the inner shell as opposed to the outer shell. The photon released is lower in energy than the x-ray emitted and has a unique identifying character allowing for identification of the element (Wirth and Barth, 2013). 60

71 Appendix C: Normalization Histograms Original well-log values Normalized well-log values GR NPHI DPHI Normalization of GR, NPHI and DPHI curves using the 5,50,95 quantile method. These are the before and after histograms of distributed values. 61

72 Appendix D: Type Log The explanation of the formation-top placement starts from the bottom and work up from there (Figure 7). The Upper Skinner Shale top (or from here on out Marmaton Base) separates the Cherokee Group and the Marmaton Group, both of which are Desmoinesian in age. As seen in the type log, the Marmaton Base separates the Marmaton Group from the Upper Skinner Shale, which has noticeably high GR, NPHI, and ILD values and relatively low DPHI values. Above the Marmaton Base top, the GR and NPHI signatures decrease significantly, and the DPHI and ILD signatures increase slightly. Marmaton F is the next top and is also known as regional flooding surface four. Thus, its placement corresponds to radioactive shale: high GR and NPHI values and low ILD and DPHI values. It is highly visible because lower GR and NPHI and higher ILD comprise the predominant surrounding signature. The Marmaton E top, also referred to as flooding surface three, contain similar signatures to the Marmaton F top. However, in this type-log, the change in NPHI, DPHI, and ILD are nominal in comparison to the increase in GR. Next stratigraphically is the Marmaton D top, or flooding surface two. Again, the overall signature for the top placement is akin to the Marmaton F top. A difference between the two in the Mayfield 1-34 log signatures shows that, instead of being juxtaposed to sandstone/conglomerate signatures above and below, above of the Marmaton D is a shaly interval that coarsens upward. The final regional flooding surface is the Marmaton C. The placement of this top in the type-log corresponds to a high GR, a high NPHI value followed by a decrease, low DPHI, and no visible change in ILD. Throughout the area of interest, the previously mentioned flooding surface tops 62

73 slightly change in character but remain true to the classic signature of a radioactive shale: high GR and NPHI with low DPHI and ILD. The remaining tops highlight shale breaks that cannot be definitively called flooding surfaces but occur regionally. Marmaton C caps a roughly coarsening upward sequence in the type log and often throughout the area. A clear increase in GR and NPHI accompany a decrease in DPHI and ILD. These signatures are a small blip in the log curves suggesting a thin shale break. The next top, Marmaton B, exhibits the same character of thin shale but in the Kula 1-33 the GR increase is innocuous. The NPHI values are high, and DPHI and ILD values are low. This suggests a gas bearing shale, which provides the guidance for the top placement. Elsewhere in the area, the GR increase is more prominent. The Marmaton A top again demonstrates the typical shale characteristic, and usually caps a coarsening upward sequence throughout the region. The Marmaton Top signals the end of the Marmaton Group. Its placement is on shale right underneath a thin sandy (or conglomeritic) package. Although on the type log it seems nonsensical to put a top at this level instead of on top of the sandy package, the character of the top interval changes throughout the basin. However, a thin signature of low GR and NPHI and elevated levels of DPHI and ILD of the sandy package exists. This affirms the reasoning for choosing this particular top for consistency. 63

74 TexasD Wheeler 64 Wells with digital or raster logs Mountain view fault C E A A B Beckham Oklahoma Roger Mills C B 8 mi 12.9 km E D N Appendix E: Regional Cross sections, Structure Maps, and Isopachs

75 Depth(ft) Wash A B C D E F A A GR NPHI and DPHI Scales Horizontal: 8203 Vertical:

76 Depth(ft) A B C D E F B B GR NPHI and DPHI Scales Horizontal: 9174 Vertical:

77 Depth(ft) Wash A B C D E F C C GR NPHI and DPHI Scales Horizontal: Vertical:

78 Depth(ft) Wash A B C D E F D D GR NPHI and DPHI Scales Horizontal: Vertical:

79 Depth(ft) Wash A B C D E F E E GR NPHI and DPHI Scales Horizontal: 1138 Vertical:

80 A B N N Thickness (ft) Wheeler County Roger Mills County Beckham County 0 mi 4 Thickness (ft) Wheeler County Roger Mills County Beckham County 0 mi km km 6.4 Marmaton Top Marmaton A C D Thickness (ft) Wheeler County Roger Mills County N Thickness (ft) Wheeler County Roger Mills County N Beckham County 0 mi Beckham County 0 mi km 6.4 Marmaton E Marmaton F Isopach maps for selected intervals in the Marmaton Group. A) The Marmaton Top interval averages thickness around 100 ft [30.5 m] with occasional areas up to 200 ft [61 m] thick. There are southwest-northeast thickness trends. B) The northern area shows that there is a thin Marmaton A interval. Directly in front of Amarillo-Wichita uplift there is a east by northeast thick. C) The Marmaton E interval has a higher average thickness than the previous two intervals. It has an expansive thin area on the northeast border of the study area with a thick occuring adjacent to the Mountain View fault system. D) The Marmaton F interval has an interesting juxtaposition of thicker sediment packages on the southwestern portion of the study area while the northeastern portion appears sediment starved km

81 B A N N Roger Mills County TVD (ft) Wheeler County 8500 Roger Mills County Wheeler County mi 4 Beckham County mi 4 Beckham County 0 km km 6.4 TVD (ft) Top of Marmaton Wash Top of Marmaton A D C N N 71 Roger Mills County Wheeler County TVD (ft) TVD (ft) Roger Mills County Wheeler County Beckham County Beckham County 0 mi mi 4 0 km km Top of Marmaton F Top of Marmaton Bottom Structure maps throughout the Marmaton Group. A) The Marmaton Wash top shows deeper elevations in the southeast corner which gradually increase the the northwest corner of the study area. There is an area adjacent to the Amarillo-Wichita uplift that shows shallower depths due to intense reverse faulting. B) The Marmaton A top exhibits a nearly identical pattern to that of the previous map. However, dirstly in front of the mountain-view fault system there is a broader much shallower section. C) The Marmaton F top lies deeper in the southeast and shallows to the northwest. D) The Marmaton Bottom Top does not have a shallow section adjacent to the uplift unlike the other tops shown. There is also a sharper increase in elevation going to the northwest in Wheeler County, suggesting the presence of a fault shown by the dashed grey line.

82 Moving in the southern direction most intervals thicken while some stay at relatively constant thickness. In cross section A-A the GR signature generally becomes cleaner (lower in values). The northern wells have prominent shale intervals and well defined sand or conglomeritic packages. To the south the packages are less defined and look similar in character throughout the Marmaton Group. There is an elevation difference of approximately 500 ft (152 m) from north to south. In cross section B-B, elevation of the Marmaton Wash drops approximately 500 ft (152 m) and packages thicken to the south as A-A did. In contrast to A-A, the two middle wells contained the poorly defined packages while the most northern and southern wells have clearly defined sand or conglomeratic packages. The end wells also have shaly periods, in particular the most southern well. The west-east cross sections also show trends in packages. In cross section C-C there is a 600 ft (183 m) change in elevation moving east and an overall increasing GR signature. Thickness remains constant until the most eastern well, which has much thinner intervals. Throughout this cross section prominent shale breaks separate sand or conglomeritic packages. Moving to the center of the study area, cross section D-D shows minimal variability within the Marmaton Group. Upper intervals (Marmaton Top, A, B, C, and D) preserve thickness moving west to east but the lower intervals (Marmaton E and F) show an increase in thickness. Well defined sand or conglomeritic packages are divided by noticeable radioactive shales. The southernmost cross section, E-E, only goes through Oklahoma and shows a different trend than the other west-east cross sections. Elevation increases from west to east approximately 300 ft (91.5 m). Intervals thin to the east and the GR values increase signaling more radioactive 72

83 elements in shale or sandstones/conglomerates. The western most well has a comparatively homogenous signature to the other wells in the cross section. To visualize the elevation changes throughout the study area, structure maps were constructed. Four formation tops were chosen to illustrate the overall trend of the upper and lower tops. The highest elevations happen in the northwest portion of the area of interest and fall in elevation to the southeast. Higher elevations also happen in the northeast corner of the area. All the formation tops follow the same deepening trend to the southern center part of the region but the area adjacent to the Amarillo-Wichita mountain front is shallower than the surrounding area. However, for the bottom most formation top no rise in elevation occurs. Trends in interval thickness are shown in isopach maps (Figure 16). Again, four tops exemplify the general tendencies of interval thickness with two examples from up section and down section. The Marmaton Top interval shows thickness trends that are perpendicular to the Amarillo-Wichita mountain front. Thinner zones and thicker zones alternate starting in the northwest moving to the southeast, with a thin area adjacent to the mountain front. Moving down section, the Marmaton A interval exhibits more of a west-east trend of thickness. A thick interval occurs next to the mountain front and thins to the north. A thin interval area occurs in the southeast of the area of interest near the Elk City high. Heading down section the thickness of the intervals dramatically increase by up to 300 ft (91.5 m). The Marmaton E interval does not appear to have any specific thickness trends. The thickest interval area is adjacent to the mountain front and the thinnest part is hugging the eastern portion of the study area. Otherwise thicknesses vary randomly by about 100 ft (30.5 m). Finally, the Marmaton F interval shows the same thinner area to the east but thick in 73

84 the southwestern portion of the area of interest. In front of the mountain front there is a thin spot of about 200 ft (61 m) surrounded by thicknesses up to 450 ft (137 m). Although minor changes occur between interval thicknesses, a constant feature is a thick area adjacent to the Amarillo-Wichita uplift. 74

85 Appendix F: Model Area Cross Sections, Structure Maps, and Isopachs C A 11N 26W 11N 25W N C B E D 10N 26W 10N 25W A E B D Wells with digital logs 3 mi 4.8 km 75

86 Depth(ft) Wash A B C D E F A A GR NPHI and DPHI Scales Horizontal: 2409 Vertical:

87 Depth(ft) Wash A B C D E F B B GR NPHI and DPHI Scales Horizontal: 1560 Vertical:

88 Depth(ft) Wash A B C D E F C C GR NPHI and DPHI Scales Horizontal: 3580 Vertical:

89 Depth(ft) Wash A B C D E F D D GR NPHI and DPHI Scales Horizontal: 2749 Vertical:

90 80 Depth(ft) Wash A B C D E F E E GR 150 Scales Horizontal: 3256 Vertical: NPHI and DPHI 0