INTEGRATED THREE DIMENSIONAL GEOLOGICAL MODEL OF THE DEVONIAN BAKKEN FORMATION ELM COULEE FIELD, WILLISTON BASIN: RICHLAND

COUNTY MONTANA

by Adrian Almanza

A thesis submitted to the Faculty and the Board of Trustees of the Colorado School of Mines in partial fulfillment of the requirements for the degree of Master of Science (Geology)

Golden, Colorado Date___________

Signed______________________ Adrian Almanza

Signed______________________ Dr. J. Frederick Sarg Thesis Advisor Golden, Colorado Date___________

Signed______________________ Dr. John D. Humphrey Professor and Head Department of Geology And Geological Engineering

ii

ABSTRACT
The Bakken Formation of the Williston basin is a oil resource play that was named the largest continuous oil accumulation in the lower 48 states by the United States Geological Survey (USGS) in 2008. Although extensive the Bakken does not have uniform properties throughout its areal extent. Identifying areas with porosity, permeability, and fractures that permit highly productive wells is essential to commercial petroleum recovery. The Elm Coulee Field is a giant oil field in Eastern Montana that exhibits some of these critical reservoir properties. Geological modeling of these reservoir properties provides a greater understanding of reservoir performance, and aids in exploration and development of the Bakken Formation. The Bakken Formation in the Elm Coulee Field consists of three members: an upper shale, middle silty dolostone, and lower shale. The Elm Coulee Oil Field is a stratigraphic trap with a pinch-out to the southwest and a diagenetic facies change in the northeast. The primary reservoir is the silty dolostone of the Middle Bakken Member. The purpose of this research is three fold: (1) complete an examination of the reservoir properties of the Bakken petroleum system in Elm Coulee field; (2) construct a three-dimensional geologic model that will show the distribution of the different facies within the Bakken Formation and their reservoir properties; and (3) build a fracture model and integrate it with the matrix porosity model. This study uses digital logs, core data, petrographic thin sections, XRD analysis, DSTs, production data, and Petrel software to characterize the Elm Coulee Field. Six cores are used to calibrate physical properties to digital well logs, and core descriptions were used to construct detailed facies maps. The study correlates the core lithofacies to digital well logs resulting in maps of the facies and the calculation of their associated thickness throughout the study area. The data was then used to build a structural geomodel in Petrel software. The geomodel is located in the central part of Elm Coulee in the congressional land blocks: Township 53 Range 56, Township 57 Range 56, and Township 53 Range 58 of Richland County, Montana. The structural model shows a thick Middle Bakken in the study area with a thinning to the north, south, and west. The Lower Bakken Shale is absent in the southern half of Richland County and the Upper Bakken Shale covers the entire study area iii

With the structural model in place, the digital well logs were used to distribute petrophysical properties of the Middle Bakken throughout the Petrel model. The data revealed that the reservoir is located in a northwest trend in the southern half of Richland County. The petrophysics shows that the best reservoir properties are associated with Middle Bakken faces B and C. Facies B and C have a high percentage of dolomite which has the best reservoir properties (i.e., higher porosity and permeability). These data were then used to build a matrix reservoir model. The study also constructed a fracture model for Richland County. The Elm Coulee fracture model was derived from seismic and production trends. The model uses regional fracture trends to establish a fracture fabric. The regional fractures are oriented to the northeast and have a spacing of approximately 1,250 feet (ft). An orthogonal set of fractures are spaced 2,500 ft apart in the northwest direction. Production data allowed for the detection of what is interpreted as fracture swarms. The swarms are oriented in the maximum principal stress direction of N60E and have an approximate spacing of 25,000 ft. The fracture model was combined with the matrix model to develop a dual porosity model. This results in a combined model that has both fracture and matrix reservoir properties that can be used in simulation. The dual porosity model reflects the production in the study area, and the best reservoir properties align with the best estimated ultimate recovery (EUR).

iv

TABLE OF CONTENTS
ABSTRACT ...................................................................................................................... iii TABLE OF CONTENTS .................................................................................................... v LIST OF FIGURES .......................................................................................................... viii LIST OF TABLES ............................................................................................................ xv ACKNOWLEDGMENTS .................................................................................................. xv CHAPTER 1 ...................................................................................................................... 1 INTRODUCTION ............................................................................................................... 1 CHAPTER 2 ...................................................................................................................... 4 GEOLOGIC OVERVIEW ................................................................................................... 4 2.1 Tectonics ..................................................................................................................... 4 2.2 Stratigraphy ................................................................................................................. 4 2.2.1 Lower Shale ............................................................................................................. 6 2.2.2 Middle Member ........................................................................................................ 7 2.2.3 Upper Shale ............................................................................................................. 7 2.2.4 Source Rocks ........................................................................................................... 8 2.3 Key Previous Work ...................................................................................................... 8 2.4 Purpose and Objectives .............................................................................................. 9 2.5 Methodology .............................................................................................................. 10 CHAPTER 3 .................................................................................................................... 11 FACIES INTERPRETATION AND CORE DESCRIPTIONS ........................................... 11 3.1 Methods .................................................................................................................... 11 3.2 Facies Descriptions ................................................................................................... 14 3.2.1 Lower Bakken Shale. ............................................................................................. 14 3.2.2 Facies A. Intraclastic-Skeletal Lime Wackestone .................................................. 14 3.2.3 Facies B. Bioturbated Silty Dolostone .................................................................... 16 3.2.4 Facies C. Rhythmic Laminated Sandy to Silty Dolostone ...................................... 16 v

3.2.5 Facies D. Fine Grained Quartz Rich Sandstone .................................................... 16 3.2.6 Facies E Silty Dolostone ........................................................................................ 16 3.2.7 Facies F Fossiliferous Wackestone ....................................................................... 16 3.2.8 Upper Bakken Shale. ............................................................................................. 21 3.3 Depositonal Model .................................................................................................... 21 3.4 Core Descriptions ...................................................................................................... 23 3.4.1 Brutus East Lewis 3-4 (Sec. 3-T24N-R57E) .......................................................... 23 3.4.2 RR Lonetree Edna 1-13 (Sec. 1-T23N-R56E) ....................................................... 26 3.4.3 Foghorn-Ervin 20-3 (Sec. 20-T23N-R58E) ............................................................ 26 3.4.3 Jackson Rowdy 3-8 (Sec. 3-T26N-R51E) .............................................................. 27 3.4.4 Vaira 44-24 (Sec. 24-T24N-R31E) ......................................................................... 28 3.4.5 Williams 1-4 (Sec. 4-T23N-R55E) .......................................................................... 29 3.5 Mineralogy and Diagenesis ....................................................................................... 29 CHAPTER 4 .................................................................................................................... 36 SUBSURFACE WELL LOG INTERPRETATION ............................................................ 36 4.1 Methods .................................................................................................................... 36 4.2 Structural Cross Section ........................................................................................... 36 4.3 Stratigraphic Cross Section ....................................................................................... 39 4.4 Structure Maps .......................................................................................................... 39 4.5 Isopach Maps ............................................................................................................ 39 CHAPTER 5 .................................................................................................................... 55 PETROPHYSICAL ANALYSIS AND RESERVOIR PROPERTIES................................. 55 5.1 Core Properties Analysis ........................................................................................... 55 5.2 Conventional Well Log Analysis ................................................................................ 56 5.3 Drill Stem Tests ......................................................................................................... 71 CHAPTER 6 .................................................................................................................... 73 FRACTURES .................................................................................................................. 73 vi

6.1 Production Data. ....................................................................................................... 73 6.2 Other Available Data ................................................................................................. 74 6.3 Seismic Mapping in Billing Nose an analog for a local fracture fabric of the Elm Coulee Field. ................................................................................................................... 74 6.4 Micro Seismic Study Elm Coulee Field. .................................................................... 78 6.5 Results ...................................................................................................................... 79 CHAPTER 7 .................................................................................................................... 83 3-D GEOLOGICAL MODELING ...................................................................................... 83 7.1 Modeling in Petrel ..................................................................................................... 83 7.3 Fracture Properties ................................................................................................... 88 CHAPTER 8 DISSCUSSION .......................................................................................... 92 8.1 Structural-Stratigraphic Model ................................................................................... 92 8.2 Porosity Model ......................................................................................................... 93 8.3 Permeability Model .................................................................................................... 93 8.4 Fractures Model ........................................................................................................ 94 8.6 Production Data ........................................................................................................ 95 CHAPTER 9 CONCLUSIONS ......................................................................................... 98 9.1 Conclusions ............................................................................................................... 98 9.2 Recommendations .................................................................................................... 99 REFERENCES CITED .................................................................................................. 100 APPENDIX A CORE DESCRIPTIONS ...................................................................... 104 APPENDIX B STRUCTURAL CROSS SECTIONS ................................................... 112 APPENDIX C STRATIGRAPHIC CROSS SECTIONS .............................................. 118 APPENDIX D LIST OF WELLS USED IN PETREL MODEL ..................................... 123

vii

LIST OF FIGURES
Figure 1.1 Location Map, Elm Coulee Field, Richland County, Montana (Heck et al.,2004). ........................................................................................................................... 1 Figure 1.2 Structure contour base of Mississippian (from Sonnenberg and Pramudito, 2009). ................................................................................................................................ 2 Figure 1.3 Digital log and core locations for Elm Coulee Field.......................................... 3 Figure 2.1 Location of major structural provinces or orogeny (Williams et al; 1991) ......... 5 Figure 2.2 Major Paleozoic structural lineaments. (Brown and Brown, 1987). .................. 5 Figure 2.3 Schematic west-east cross section across the Williston basin (from Sonnenberg and Pramudito (2009) after Meissner, (1978)).............................................. 6 Figure 2.4 Well log signatures of the Bakken Formation. (from Sonnenberg and Pramudito., 2009). ............................................................................................................. 7 Figure 3.1 Green dots mark the locations of cored wells used in the Elm Coulee study. 13 Figure 3.2 Bakken lithofacies. ......................................................................................... 15 Figure 3.3 Lower Bakken Shale. The black and white scale bar segments are one inch (from Alexandre, 2011). ........................................................................................... 17 Figure 3.4 Middle Bakken Facies A. The black and white scale bar segments are one inch (from Alexandre, 2011). ........................................................................................... 17 Figure 3.5 Middle Bakken Facies B. The black and white scale bar segments are one inch. Dark blebs are Helminthopsis burrows (from Alexandre, 2011). ............................ 18 Figure 3.6 Middle Bakken Facies C. The black and white scale bar segments are one inch (from Alexandre, 2011). ........................................................................................... 18 Figure 3.7 Middle Bakken Facies D. The black and white scale bar segments are one inch (from Alexandre, 2011). ........................................................................................... 19 Figure 3.8 Middle Bakken Facies E. The black and white scale bar segments are one inch (from Alexandre, 2011). ........................................................................................... 19 Figure 3.9 Middle Bakken Facies F. The black and white scale bar segments are one inch (from Alexandre, 2011). ........................................................................................... 20 Figure 3.10 Upper Bakken Shale. The black and white scale bar segments are one inch (from Alexandre, 2011). ........................................................................................... 20 Figure 3.11 North American Paleogeographic Map Late Devonian, 360 Ma (Modified from Blakey 2005)..22

viii

Figure 3.12 Bakken depositional model. Facies are labeled according to their position on the model. HST (highstand systems tract), LST (Lowstand systems tract), TST (transgressive systems tract)(from Simenson, 2010 modified from Smith and Bustin, 1996). .............................................................................................................................. 23 Figure 3.13 Brutus East Lewis 3-4-H core description (from Alexandre, 2011) .............. 25 Figure 3.14 RR Lonetree, gamma ray, porosity, and permeability and XRD analysis. ... 31 Figure 3.15 Jackson Rowdy, gamma ray, porosity, and permeability and XRD analysis. .......................................................................................................................... 32 Figure 3.16 Brutus East Lewis cross plot of bulk density and XRD precentage of dolomite and calcite (limestone). The colors refer to the depth of the Middle Bakken. The red is at 7,636 ft and the dark blue is 7,6578 ft. Data shows that a high precentage of dolomite corrospons to a bulk density of 2.63 g/cm3 and has a calicte signature of 2.68 g/cm3 . ................................................................................................. 33 Figure 3.17 RR Lonetree Edna, cross plot of bulk density and XRD precentage of dolomite. The colors refer to the depth of the Middle Bakken. The red is at 10,376 ft and the dark blue is 10,408 ft. The data shows an average bulkdensity of 2.63 g/cm3 where there is a high precentage of dolomite. ................................................................ 34 Figure 3.18 Average bulk density map of the Middle Bakken. The average dolomite signature (2.63-2.66 g/cm3) is observed in the southern half of the study area. The calcite signature 2.68-2.69) g/cm3 is seen in the northern and western parts of the study area and are considered diagenetic pinch outs of reservoir properties. ................ 35 Figure 4.1 Locatoin map of digital wells. Yellow squares mark digital well locations...37 Figure 4.2 Location map of structural and stratigraphic cross sections. Top Bakken structure map contour interval is 200 ft..38 Figure 4.3 Structural cross section G-G'. The Bakken (purple), Three Forks (Green), and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black)...40 Figure 4.4 Stratigraphic corss section G-G'. The dark gray is the Upper and Lower Bakken shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A.41

ix

Figure 4.5 Thickness map of the Middle Bakken Member (facies A,B,C,D,E, and F). Contour interval is 2.5 feet. The thickness ranges from 5.5 feet to 52 feet in the study area. The Middle Bakken Member thins to the north and east and pinches out to the south and west. ......................................................................................................... ..42 Figure 4.6 Structure Map Upper Bakken Shale. The structure map shows a dip to the east. The contours interval is 100 ft. ............................................................................... 44 Figure 4.7 Structure Map Three Forks. The structure map shows a dip to the east. The contours interval is 100 ft. ............................................................................................... 45 Figure 4.8 Structure Map Nisku (Birdbear) The structure map shows a dip to the east. The contours interval is 100 ft. ........................................................................................ 46 Figure 4.9 Thickness map of the Upper Bakken Shale. Contour interval is one foot (0.30 meters). The thickness ranges from 4.5ft to 13ft in the study area. ...................... 47 Figure 4.10 Thickness map of the CSM F facies. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to five feet in the study area.......................... 48 Figure 4.11 Thickness map of the CSM E facies. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to 7.5 ft in the study area .............................. 49 Figure 4.12 Thickness map of the CSM C facies. Contour interval is two feet (0.61 meters). The thickness ranges from 2.4ft to 17ft in the study area ................................. 50 Figure 4.13 Thickness map of the CSM B facies. Contour interval is one foot (0.30 meters). The thickness ranges from seven feet to 27 ft in the study area ....................... 51 Figure 4.14 Thickness map of the CSM A facies. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to eight feet in the study area ....................... 52 Figure 4.15 Thickness map of the Lower Bakken Shale. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to six feet in the study area .................. 53 Figure 4.16 Thickness map of the Three Forks. Contour interval is ten feet (three meters). The thickness ranges from 1.5ft to five feet in the study area........................... 54 Figure 5.1 Porosity and Permeability cross plots. The Foghorn and Lonetree Wells have good logarithmic regression line correlation. The Brutus and the Jackson Rowdy well have poor logarithmic line colorations. The former two wells are situated in a dolomitic setting as indicated by bulk density and XRD analysis. ................................... 57 Figure 5.2 Middle Bakken porosity and permeability crossplot (from Enerplus core data) ................................................................................................................................ 58 Figure 5.3 Photomicrograph showing examples of nonreservoir (A) and good reservoir (B)..60 x

Figure 5.4 Foghorn Ervin Core Data: gamma ray (GR), bulkdensity (ROBH), neutron porosity (NPHI), water saturatoin (Core Sw), oil saturation (Core So), Porosity, Permeavility(Core Kmax). Log templates constructed in Prism module of Geographix...61 Figure 5.5 Brutus East-Lewis Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core so), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix. ..................................................................................................................... 62 Figure 5.6 Lonetree-Edna Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core so), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix. ..................................................................................................................... 63 Figure 5.7 Jackson Rowdy Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core so), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix. ..................................................................................................................... 64 Figure 5.8 Viara Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core so), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix. ................... 65 Figure 5.9 Williams Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core so), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix. ..................................................................................................................... 66 Figure 5.10 Facies C and B average porosity map of the Elm Coulee Field. The Best porosity is found in the southern half of the study area........67 Figure 5.11 Jackson Rowdy, the thick red lines in the Porosity and PERM tracks are the calculated average porosity (PHIA) and permeability (Perm). ................................. 68 Figure 5.12 Foghorn Ervin, the thick red lines in the Porosity and PERM tracks are the calculated average porosity (PHIA) and permeability (Perm). ................................. 69 Figure 5.13 Middle Bakken SoPhiH Map indicating most petrophysicaly prospecitve areas of Elm Coulee field...70 Figure 5.14 Horner plots of Elm Coulee Wells ................................................................ 72 Figure 6.1 First year cumulative production .................................................................... 75 xi

Figure 6.2 First year cumulative production and Operators ............................................ 75 Figure 6.3 Initial production trends with a N60E Bias. This bias data trend highlights the fracture swarms that are interpreted in the study area .............................................. 76 Figure 6.4 Minimum curvature attirbute map, Bakken horizon (from Angster 2010). Interpreted fractures are on the right map....77 Figure 6.5 Summery of micro seismic observations (from O'Brien, Larson, & Parham, 2011...78 Figure 6.6 Micro seismic production traces and fracture height data. The average fracture height was 290 ft. Radioactive tracer show zonal isolation averaged approximately 1200ft (O'Brien, Larson, & Parham, 2011). .............................................. 79 Figure 6.7 Fracture lengths versus fracture spacing. The logarithmic plot shows the power log relationship between the production, seismic and micoseismic data used in the fracture model. ...................................................................................................... 81 Figure 6.8 Conceptual fracture model. The fracture swarm trends in the blue ovals and are spaced ~4.75 miles or 25,000ft apart. The swarms have the greatest influence on the fracture fabric. The green lines reflect the regional fracture fabric that are oriented with the maximum principal stress and have a spacing of ~1,250ft this was seen in both the micro seismic and Bicentennial seismic survey. The oranges lines show the orthogonal spacing of fractures that are spaced ~2,500ft apart. ............. 82 Figure 7.1 The red box is the outline of the Petrel model and shows the location of the digital wells used to distribute petrophysical properties. ........................................... 84 Figure 7.2 Petrel geo-model structural model and geo-cellar grid. The mapped horizons are the Three Forks, Bakken, and Lodgepole Formations. .............................. 85 Figure 7.3 Middle bakken porosity distributioin. 3D view looking north. Purple is low (0.0%) and red is high (20.0%).....86 Figure 7.4 Middle Bakken permeability distribution. 3D view looking north. Purple is low (0.001mD) and red is high (0.1mD) log scale ........................................................... 87 Figure 7.5 Localized fractures derived from Centennial seismic and the micro-seismic study from Enerplus. The dominant fracture trend is the northeast orientations (blue polygons) and the orthogonal northwest fracture set is represented by the purple polygons. ......................................................................................................................... 88

xii

Figure 7.6 The figure shows the relationship of the two fracture patterns 1. regional fracture pattern (purple polygons) and 2. contoured first year cumulative production data, cropped from figure 6.3. The contoured interval is 2000 barrels of oil. Orange is 18,000 barrels and the blue is 4,000 barrels of oil.89 Figure 7.7 Localized fractures derived from Centennial seismic and the mico-seismic study from Enerplus. The fractures have been related to the first year cumulative production data and have varying degrees of fracture influence....90 Figure 7.8 Oblique view highlighting the fracture variability in the fracture model. The blue lines are the northwest fractures and the red lines area the northeast fractures....91 Figure 7.9 Dual porosity model. Model shows the matrix model combined with the fracture model creating a dual porosity model. ............................................................... 91 Figure 8.1 Diagram of the integrated geomodeling workflow showing the different component models .......................................................................................................... 93 Figure 8.2 Initial first year cumulative production vs. Operator. The data does not show any definite production trends associated with individual operator.94 Figure 8.3 EUR well data. The graph shows the EUR production trends (>6%porosity+ fractures, >6%porosity-fractures,<6%porosity. The P90 and P10 were calculated by fitting a regression line to the data trends and finding the intercept that corresponded 1.28 standard distribution...96 Figure 8.4 EUR well data. The table shows the difference between and matrix and sweet spot (matrix and fracture trends) production. Mean production shows average production in the trends..96 Figure 8.5 Fracture trends and bubble map with calculated EUR production........97 Figure A1 Summery of Bakken Formation lithofaces used in figures A4-A8. (from Alexandre, 2011) ........................................................................................................... 104 Figure A2 Legend of symbols used in the interpretations and measured sections of figures A4-A6. This bioturbation index was first used by Reineck (1963) (from Alexandre, 2011) ........................................................................................................... 105 Figure A3 Legend of symbols used in the interpretations and measured sections of figures A7-A8 (from Pramudito, 2008) .......................................................................... 106 xiii

Figure A4 RR Lonetree-Edna 1-13 Core description (from Alexandre, 2011)............... 107 Figure A5 Foghorn-Ervin 20-3-H Core description (from Alexandre, 2011) .................. 108 Figure A6 Jackson Rowdy 3-8 Core description (from Alexandre, 2011) ..................... 109 Figure A7 Vaira 44-24 Core description (modified from Pramudito, 2008) ................... 110 Figure A8 Williams 1-4 Core description (modified from Pramudito 2008) ................... 111 Figure B1 Structural cross sections A-A`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). ................. 113 Figure B2 Structural cross sections B-B`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). ................. 114 Figure B3 Structural cross sections C-C`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). ................. 115 Figure B4 Structural cross sections D-D`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). ................ 115 Figure B5 Structural cross sections E-E`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). ................. 116 Figure B6 Structural cross sections F-F`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). ................. 116 Figure B7 Structural cross sections H-H`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black) .................. 117 Figure C1 Stratigraphic cross sections A-A`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A. .................................................................................................... 119 Figure C2 Stratigraphic cross sections B-B. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A. .................................................................................................... 119

xiv

Figure C3 Stratigraphic cross sections C-C`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and purple is facies A. .......................................................................................................... 120 Figure C4 Stratigraphic cross sections D-D`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and purple is facies A. .......................................................................................................... 120 Figure C5 Stratigraphic cross sections E-E`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and purple is facies A. .......................................................................................................... 121 Figure C6 Stratigraphic cross sections F-F. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and purple is facies A. .......................................................................................................... 121 Figure C7 Stratigraphic cross sections H-H`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and purple is facies A. .......................................................................................................... 122

LIST OF TABLES

Table 1. List of cored wells used in study. ...................................................................... 12 Table 2. Summary of service company DST results, compared to average core permeability data (Sorenson well is in close proximity to the foghorn well and, Bahl well is close to the Lonetree well.71

xv

ACKNOWLEDGMENTS

I would like to express my deep gratitude for all those that have made this possible. First and foremost I would like to thanks my parents who have been my biggest supporters and have provided me with love and encouragement to live to my best potential. I would like give thanks to my advisor Rick Sarg and my committee members Steve Sonnenberg and Hossein Kazemi for all their invaluable impute and mentorship. I like to thank NETEL for providing funding for my thesis. Finally I would like to thank my fellow cohorts of geology whose struggles, triumphs and most importantly friendships helped my own journey.

xvi

CHAPTER 1 INTRODUCTION
The Devonian-Mississippian (~350 million ma) Bakken Formation is the largest continuous US oil accumulation assessed by the USGS. The Elm Coulee field is one of the first discovered fields and currently produces over 50,000 barrels of oil per day (Figure 1.1). The Bakken Formation consists of three members: an upper shale, middle silty dolomite, and lower shale. The middle member is the main reservoir and consists of six identified facies, from the base of the member upward: facies A,B,C,D,E, and F. The two primary reservoir facies are B and C. They are characterized by a bioturbated silty dolostone and a laminated silty dolostone respectively. The Bakken reservoir rocks are characterized overall by low porosity and low permeability.

Figure 1.1. Location Map, Elm Coulee Field, Richland County, Montana (Heck et al.,2004).

Elm Coulee is located in Richland County, Montana in the southwest part of the Williston basin (Figure 1.2). In the study area, the Middle Bakken interval is the thickest in the middle of Richland County, and this thick trends in a northwest and southeast direction. This unit thins in the northeast and southwest directions. In the study area, there are six wells that have core data, and approximately 400 wells with digital logs. The digital logs have good areal coverage in the field area, but are sparser in the northeast portion of the field (Figure 1.3).

Figure 1.2. Structure contour base of Mississippian (from Sonnenberg and Pramudito, 2009).

Figure 1.3. Digital log and core locations for Elm Coulee Field.

CHAPTER 2

GEOLOGIC OVERVIEW
2.1 Tectonics
The Bakken Formation was deposited in the intracratonic Williston basin, which formed in the Late Cambrian time (~500 ma). The Williston basin is thought to have been formed by two major tectonic events. The first event was Precambrian convergence of the Churchill Hinterland and the Superior Province, which formed the Trans-Hudson Orogen (Figure 2.1). This newly created province was then subject to uplift and erosion from thermal expansion, resulting in three kilometers of overburden removal (Crowely, 1985). In the late Precambrian (~1000 ma), rifting of the ancient North American plate created a series of major basement faults that now control the structures and geometries of the Williston basin (Gerhard, 1990). The second major tectonic event occurred during Late Paleozoic (~300 ma) (Antler and Ancestral Rocky Mountains orogenies), in which plate collision from the west realigned the deformation, and reactivated these deep-seated lineaments (Figure 2.2). These were overprinted by the Laramide Orogeny in the latest Cretaceous (~60 ma) and early Tertiary forming the structural features seen in the basin today.

2.2 Stratigraphy
The total preserved stratigraphic fill of the Williston basin is approximately 16,000 feet (ft). The sediment infill is characterized by cyclic deposition of clastics and carbonates that range in age from Cambrian to Quaternary. Subsidence and tectonic basin reconfiguration are the primary influences facilitating sediment infill. The Cambrian strata are the result of Paleozoic seas that transgressed over the Williston Basin, depositing sandstones, shales and shallow water carbonates. The Cambrian is overlain by an unconformity that represents a sea level drop. The Ordovician to Silurian strata mirror the basin geometry and are characterized by carbonates and anhydrites (Vigrass, 1971).

Figure 2.1. Location of major structural provinces or orogeny (Williams et al; 1991)

Figure 2.2. Major Paleozoic structural lineaments. (Brown and Brown, 1987).

From Silurian to Devonian time, structural movement along the Trans-Continental Arch caused a major unconformity and a seaway reorientation to the north. The Devonian-Mississippian strata were deposited in this restricted basin and are characterized by transgressive-regressive cycles. The strata that were deposited are composed of carbonates, evaporites, and organic-rich shales. At the end of Mississippian time (~320 ma), the basin experienced an additional reorientation that occurred from the west. The focus of this study is the Bakken Formation, which was deposited from Late Devonian to early Mississippian time (~374-330 ma). In Montana, the Bakken is uncomformably deposited on top of the Devonian Three Forks Formation. The Bakken consists of three members: the lower member is an organic-rich shale; the middle member is a silty dolostone; and the uppermost member is an organic-rich shale (Figure 2.4). The Bakken is overlain by the Lodgepole Formation. The Lower and Upper Bakken Shales are considered source rocks for the Middle Bakken Member, Lodgepole Formation, and Three Forks Formation. The Elm Coulee field forms a stratigraphic trap by onlap of the middle Bakken strata to the northwest, northeast, and southwest (Sonnenberg and Pramudito, 2009) (Figure 2.3).

Figure 2.3. Schematic west-east cross section across the Williston basin (from Sonnenberg and Pramudito (2009) after Meissner, (1978)). 2.2.1 Lower Shale The lower shale ranges from siltstone to organic rich mudstone, and is dark gray to brown in color. The shale is burrowed and contains fossil fragments. The lower shale changes laterally from high resistivity and high gamma ray shale in the north, to lower resistivity and lower gamma ray siltstone in the south (Sonnenberg and Pramudito,

2009) (Figure 2.4). The lower Bakken Shale is areally limited and pinches out to the west and southwest in Elm Coulee Field (Figure 2.3). 2.2.2 Middle Member The Middle Bakken Member is gray to buff, dolomitic fine-grained siltstone. The middle member has ripple laminations, planar laminations, bioturbation and soft sediment deformation. This interval is interpreted to be deposited in a tide-dominated coastal environment. The middle member has a distinctive log signature (Figure 2.4). The upper silty dolostone facies has a clean, low gamma-ray signature and the bioturbated dolostone facies has a coarsening-upward log signature (Sonnenberg and Pramudito., 2009). 2.2.3 Upper Shale The upper shale is an organic rich, finely laminated, black to fissile, and slightly calcareous source rock. The upper shale has disseminated kerogen throughout, and shows a slight thickening toward the northwest. High gamma ray and resistivity signatures are characteristic of this shale (Figure 2.4).

Figure 2.4. Well log signatures of the Bakken Formation. (from Sonnenberg and Pramudito., 2009).

2.2.4 Source Rocks

These shales are thermally mature at varying depths in the Williston basin. In those places where the geothermal gradients were highest, hydrocarbon generation occurred at depths of about 7,700 ft; and in areas with lower geothermal gradients, generation occurred at depths of 10,000 ft (Price et al., 1984). Based on conodont color alteration index values, Hayes (1985) also confirmed that the Bakken shales were capable of oil generation at depths between 7,500 and 10,000 ft. Though the Bakken was rapidly buried during the Mississippian, subsidence was not enough to initiate oil generation until about 100 million years ago during the Cretaceous. Then in the later part of the Cretaceous, once oil generation began and pressures built up, a continuous phase oil expulsion occurred (Dow, 1974; Webster, 1984; Nordeng, 2009).

2.3 Key Previous Work

Petroleum geology of the giant Elm Coulee Field, Williston Basin (2009), by Stephen Sonnenberg and Airs Pramudito, is a comprehensive summary of the Elm Coulee field, and describes the stratigraphy, organic richness of the shale, and properties of the Middle Bakken Member. Furthermore, it analyzes the porosity and permeability of the middle member, and the distribution of the facies in the Elm Coulee Field area. The paper also explains that the thickening of the Bakken in the Elm Coulee area is a direct result of salt dissolution from the Prairie Formation, which created extra accommodation space for the thickening of the Middle Bakken member in Richland County. Late Devonian and Early Mississippian Bakken and Exshaw Black Shale Source Rocks, Western Canada Sedimentary Basin: A Sequence Stratigraphic Interpretation, by Mark Smith and Marc Bustin (2000). The authors state that the Bakken and Exshaw formations are time equivalent petroleum systems. They interpret three distinctive system tracts: a transgressive systems tract, a lowstand systems tract, and a second trangressive systems tract for the Middle Bakken member. They also interpret a sequence boundary between the lower and middle member of the Bakken Formation. They describe the Middle Bakken as a shallow marine deposit 8

that is characterized by an average upward increase in grain size, upward thickening in sandstone beds, and an upward decrease in bioturbation. Petroleum Geology of the Bakken Formation Williston Basin, North Dakota and Montana, by Fred Meissner (1978). Meissner discusses the maturity of the Bakken Formation and its relation to petrophysical properties. Meissner also examines the overpressure in the Williston basin, and makes an interpretation on the origin and migration of hydrocarbons from overpressured cells. Meissner states that areas with mature oil-generating source rocks generally are overpressured, which has the potential to fracture host and or adjacent rocks.

2.4 Purpose and Objectives

The purpose of this research is three fold: (1) to complete an in-depth examination of the reservoir properties of the Bakken petroleum system in Elm Coulee field; (2) to construct a three dimensional geologic model that will show the distribution of the different facies within the Bakken Formation and their associated reservoir properties: and (3) build a fracture model and integrate it with the matrix porosity model. Since porosity and permeability are highly dependent on the reservoir facies distribution, the knowledge of such properties will be useful in locating drilling targets, and in predicting the most efficient way to develop the field. Porosity can be both matrix and fracture in nature, and in Elm Coulee, matrix porosity and permeability, enhanced by natural fractures, will be key in identification of reservoir sweet spots. Furthermore, the construction of a reservoir model will facilitate identifying key production areas, and aid in forecasting and developing secondary recovery techniques. To maximize the potential of the Elm Coulee Field, a proper understanding of trap, seal and reservoir heterogeneity is critical. Ultimately, the findings of this research may be used as an analog for similar Bakken fields and play types that produce from tight reservoirs. The specific plans of this research are to: 1) Understand the lateral and vertical facies variation within the Bakken Formation in the Elm Coulee Field; 2) Correlate core data to the digital log data and distribute these properties throughout the field;

3) Build a reservoir model that displays these properties, and provides information on the depositional history of the Elm Coulee Field; 4) Use production data, drill stem tests, micro-seismic, 3D seismic and published literature to identify fractures in the Elm Coulee Field. Key producing areas in Elm Coulee Field can be found using traditional reservoir properties, and sweet spots can be identified using production trends. It is hypothesized that the best production is associated with fractured areas. This study will build both a matrix and fracture distribution model.

2.5 Methodology
The research was conducted using core data, X-ray diffraction (XRD) data, and digital log data. The geologic model was built using the geologic software: Petrel, Geographix, and Prism log analysis. The study has six cores that were used to identify facies, fractures, and reservoir properties (Alexandre, 2011). These cores where then used to correlate approximately 400 digital logs in the Elm Coulee area. The cores were also used to identify potential baffles and barriers that contribute to the heterogeneity of the Elm Coulee reservoir facies. The well logs were used to construct structure and isopach maps in the Elm Coulee Field, as well as to identify petrophysical properties. The well log data will all be spatially referenced and depth calibrated to be imported into Petrel 3D reservoir modeling software. The reservoir model focuses on the Middle Bakken Member in the Elm Coulee Field, and the core properties were used to construct a reservoir model in Petrel. This model identifies the distribution of facies and their reservoir properties, as well as the vertical and lateral changes in the Middle Bakken Member. Reservoir pressure was determined from drill stem test data in the study area. Production data from IHS was evaluated to help identity the best producing areas, as well as to assist in history matching and fluid recovery rates for the Petrel model.

10

CHAPTER 3

FACIES INTERPRETATION AND CORE DESCRIPTIONS

Richland County has a number of vertical conventional cores through the Middle Bakken, but only six cores were available for this study. Two of the cores were previously studied and interpreted by Pramudito (2009) and four new cores were interpreted and added to the CSM Elm Coulee database (Alexandre, 2011). The core data is spatially located across the study area (Figure 3.1). The core data was used to determine facies distributions and rock properties. The core dataset for this study includes four primary cores in Richland County that were donated by Enerplus Oil and Gas Corporation (Jackson Rowdy 3-8, Township 26 North, Range 51 East, Section 3; Brutus East Lewis 3-4-H, Township 24 North, Range 57 East, Section 3; Foghorn-Ervin 20-3-HILD3, Township 23 North, Range 58 East, Section 20; RR Lonetree-Edna 1-13, Township 23 North, Range 56 East, Section 1). Two other cores were used to compliment and compare to the Enerplus core data (Vaira 44-24, Township 24 North, Range 54 East, Section 24; Williams 1-4, Township 23 North, Range 55 East, Section 4). These two cores are located at the United States Geological Survey (USGS) Core Research Facility in Denver, Colorado. (Table 1) shows a summary of the cores used in this study. The cores were studied and used to identify facies and their associated properties such as porosity and permeability. Each Enerplus core has approximately 30 core plugs taken at one foot intervals. The plugs are used to determine porosity and permeability. The core plugs were thin sectioned and had X-ray diffraction (XRD) analysis run on them. The supporting data was then used to quantify grain texture, fractures, digenesis and mineral composition.

3.1 Methods
Cores were described (Alexandre, 2011) (Pramudito, 2008) for their lithologic characteristics, variations in mineral composition, grain size, fossil and trace fossil associations, diagenetic features and sedimentary structures. Grain size was measured using hand lens and petrographic thin sections.

11

12 Table 1. List of cored wells used in study.

Figure 3.1. Green dots mark the locations of cored wells used in the Elm Coulee study.

Petrographic thin sections and XRD data were used to determine mineral composition. Sedimentary structures such as laminations, lag deposits and disturbed bedding helped determine depositional environment. The presence of calcite and dolomite minerals was determined with the use of XRD data and petrographic thin sections. The presence of dolomite could also be observed in thin sections in the form of dolomite crystals with rhombic shapes. XRD analysis measured the mineral percentages contained in the middle member. The Middle Bakken facies B and C with a high percentages of dolomite had the best reservoir rock. To map the facies and paleoenviormental interpretations, core data was correlated to well log data. The well log data was used to construct cross-sections across the study area. A series of structure and isopach maps were constructed to display the areal distributions and thicknesses of the Middle Bakken facies. 13

3.2 Facies Descriptions

A facies is a body of rock characterized by a particular combination of lithology, physical and biological structures that exhibit an aspect different from the bodies of rock above, below, and laterally adjacent (Walker, 2006). Core observation, petrographic thin sections, and XRD were utilized to help with facies interpretation. One of the most important features for facies identification was fossil identification, and both trace fossils and fossil fragments were instrumental in determining facies boundaries (Alexandre, 2011). The six cores used in this study all exhibit similar facies characteristics and petrophysical properties. The Middle Bakken interval has six distinct facies that were identified based on sedimentary structures, fossil content, textures and mineral composition. The Middle Bakken has a facies designation from the base upwards of lithofacies A,B,C,D,E, and F (Figure 3.2) (CSM 2010). These facies are summarized in the following sections. The identified facies correlate well with gamma ray, neutron and density log properties. The average core plug data from Enerplus was used for matrix properties values. 3.2.1 Lower Bakken Shale. The lower shale is an organic rich mudstone, and is dark gray to brown in color. The shale is burrowed and contains fossil fragments including brachiopods and crinoids (Figure 3.3). The lower shale is areally limited and pinches out to the west and southwest in Richland County. Thin sections reveal that this facies has micro fractures that are generally cemented by calcite and pyrite. Core data reports indicate an average porosity of 3.1%. 3.2.2 Facies A. Intraclastic-Skeletal Lime Wackestone Lithofacies A is composed of a silty clay-rich wackestone (Figure 3.4). The facies is characterized by abundant fossil fragments consisting of brachiopods and crinoids. Some cores have a low to moderate amount of bioturbation in the form of horizontal burrows. Pyrite is present in the form of wisps and nodules (Alexandre, 2011). Petrophysical logs show an increase in

14

15 Figure 3.2. Bakken lithofacies.

gamma ray values that generally are assumed to be associated with an increase in organic material (Pramudito, 2008). The thickness of the interval averages four feet. The average porosity is four percent and average permeability is 0.12 md. 3.2.3 Facies B. Bioturbated Silty Dolostone Lithofacies B is composed of a highly bioturbated silty dolostone (Figure 3.5). The facies is characterized by a high amount of bioturbation. The primary trace fossils are helminthopsis burrows. There are rare occurrences of thin laminations and brachiopod fragments. The clay content decreases toward the top of this interval. The thickness of the interval averages 15 ft. The average porosity is six percent and the average permeability is 0.1 md. 3.2.4 Facies C. Rhythmic Laminated Sandy to Silty Dolostone Lithofacies C is composed of a laminated silty dolostone (Figure 3.6). The facies is characterized by the rhythmically bedded laminations. The types of lamination observed in the facies are parallel, wavy, and cross laminations. There is a low amount a bioturbation observed in this facies. The thickness of the interval averages seven feet. The average porosity is six percent and the average permeability is 0.06 md. 3.2.5 Facies D. Fine Grained Quartz Rich Sandstone Lithofacies D is composed of dolomitic fine-grained sandstone (Figure 3.7). The types of lamination observed in the facies are parallel, wavy, and cross laminated. The facies is very thin, averaging less than one foot. This is below the resolution of the digital curves and is not mapped in the field area. This facies is combined with facies C when mapping across the study area. 3.2.6 Facies E Silty Dolostone Lithofacies E is composed of a laminated to bioturbated silty dolostone (Figure 3.8). There are wavy discontinuous laminations observed throughout this interval. The thickness of the interval averages 2.6 ft. This is not a major reservoir facies in the Middle Bakken. 3.2.7 Facies F Fossiliferous Wackestone Lithofacies F is composed of fossiliferous dolomitic to lime wackestone (Figure 3.9).

16

Figure 3.3. Lower Bakken Shale. The black and white scale bar segments are one inch (from Alexandre, 2011).

Figure 3.4. Middle Bakken Facies A. The black and white scale bar segments are one inch (from Alexandre, 2011).

17

Figure 3.5. Middle Bakken Facies B. The black and white scale bar segments are one inch. Dark blebs are Helminthopsis burrows (from Alexandre, 2011).

Figure 3.6. Middle Bakken Facies C. The black and white scale bar segments are one inch (from Alexandre, 2011).

18

Figure 3.7. Middle Bakken Facies D. The black and white scale bar segments are one inch (from Alexandre, 2011).

Figure 3.8. Middle Bakken Facies E. The black and white scale bar segments are one inch (from Alexandre, 2011).

19

Figure 3.9. Middle Bakken Facies F. The black and white scale bar segments are one inch (from Alexandre, 2011).

Figure 3.10. Upper Bakken Shale. The black and white scale bar segments are one inch (from Alexandre, 2011).

20

The beds contain brachiopods found as articulated shells or fragments. There is fair to moderate bioturbation observed. The thickness of the interval averages 2.6 ft. The facies is generally hard to detect on wire line logs do to its thin nature and its proximity to the high radioactivity of the overlying Upper Bakken Shale. 3.2.8 Upper Bakken Shale. The upper shale is an organic rich mudstone, and is black to dark brown in color (Figure 3.10). The Upper Bakken Shale is areally distributed throughout the Elm Coulee Field. The facies has micro fractures that are generally cemented by calcite and pyrite. The upper shale has disseminated kerogen throughout, and shows a slight thickening toward the northwest. High gamma ray and resistivity signatures are characteristic of this shale. Core measurements show an average porosity of 1.7%. 3.3 Depositonal Model The Bakken Formation was deposited in a restricted basin during a time of relative sea level rise (Webster,1984; Price,1985) (Figure 3.11). The lower and upper organic rich shales are interpreted to be deposited in a stratified water column. The anoxic bottom waters allowed for the preservation of high amounts of organic matter, an absence of bioturbation and the formation of pyrite. (Webster,1984; Price,1985). The depositional model for the Middle Bakken in Richland County is an offshore marine carbonate bar complex (Sonnenberg and Pramudito, 2009) The carbonate bar complex is interpreted to be deposited in accommodation that was the result of the dissolution of the underlying Prairie salts (Sonnenberg and Pramudito, 2009) Facies A,B, and C are interpreted to have been deposited in a highstand systems tract (HST) (Figure 3.12). Facies A is interpreted to be deposited in an offshore marine environment. The facies does not have any wave or current features. There is a very low amount of bioturbation, Facies B is interpreted to be have been deposited in a lower shoreface environment. This facies has a high amount of bioturbation from deposit feeders and shows little to no wave or current features (Simenson, 2010). Facies C is interpreted to have been deposited in the intertidal environment, and is characterized by laminated silty dolostone. The top of Facies C marks the end of the HST. Facies E is interpreted to have been deposited in a lower intertidal environment. There is fair to moderate amount of bioturbation and some current and tidal features.

21

Facies F is interpreted to have been deposited in an offshore subtidal environment due to an increase in clay content. There are also abundant fossil fragments concentrated in thin layers, that are interpreted to be storm deposits. Facies E and F are interpreted to have been deposited during a transgressive systems tract (TST) (Figure 3.12).

Williston Basin

Figure 3.11 North American Paleogeographic Map Late Devonian, 360 Ma (modified from Blakey 2005).

22

Figure 3.12. Bakken depositional model. Facies are labeled according to their position on the model. HST (highstand systems tract), LST (Lowstand systems tract), TST (transgressive systems tract)(from Simenson, 2010 modified from Smith and Bustin, 1996). 3.4 Core Descriptions Cores were examined using hand lens, petrographic thin sections and aided by the use of digital well logs, XRD and core plug sample analysis. Two core wells, (Vaira, Williams), were previously studied by Pramudito (2009). A comprehensive analysis of the four Enerplus Oil Corporation cores has been completed Alexandre (2011), and the following descriptions are a brief summary of this work. The facies found in these key well cores were used to define layers in the geologic model. 3.4.1 Brutus East Lewis 3-4 (Sec. 3-T24N-R57E) The Brutus Lewis 3-4 well is located in the north-east area of the Elm Coulee Field. The cored interval is 10,372-10,435 ft and contains both of the Bakken Shales and six of the Bakken facies: A, B, C, D, E, and F (Figure 3.13).

23

The lower shale is an organic rich mudstone that is dark brown in color. The shale has thin millimeter sized laminations. The facies has microfractures that are generally cemented by calcite and pyrite. Facies A is composed of a silty clay rich wackestone that is light gray to gray. There are fossil fragments that consist of brachiopod and crinoid fragments. There is a fair to moderate amount of bioturbation in the form of horizontal burrows. Pyrite is present in the form of wisps and nodules. Facies B is composed of a silty dolostone that is gray to light brown. Facies B has a high amount of bioturbation. The primary trace fossils are helminthopsis burrows. There are rare occurrences of thin laminations and brachiopod fragments. Facies C is composed of a laminated silty dolostone that is light gray to light brown. Facies C is characterized by the rhythmically bedded laminations. The types of lamination observed in the facies are parallel, wavy, and cross laminated. There is a low amount a bioturbation observed in this facies. Facies D is very fine grained sandstone that is light brown. This facies is quartz rich. There are millimeter sized laminations with some minor amount of bioturbation found in Facies D. Facies E is composed of a laminated to bioturbated silty dolostone that is gray to light brown. There are wavy discontinuous laminations observed throughout this interval. Facies F is a fossiliferous dolomitic to lime wackestone that is dark gray to brown. There are sporadic laminations found through the section, and fair to moderate bioturbation is observed. The upper shale is an organic rich mudstone, and is black to dark brown in color. The shale has millimeter sized laminations, is slightly burrowed, and contains minor fossil fragments. The upper contact is sharp with the overlying Lodgepole Formation.

24

Figure 3.13. Brutus East Lewis 3-4-H core description (from Alexandre, 2011) 25

3.4.2 RR Lonetree Edna 1-13 (Sec. 1-T23N-R56E) The Lonetree Edna well is located just east of the central part of Elm Coulee Field. The cored interval is 10,362-10,421 ft and contains only the Upper Bakken Shale and three of the Bakken facies: A, B, and C (Figure A4). Facies A is composed of a silty clay-rich wackestone that is light gray to gray. The facies has abundant fossil fragments that consist of brachiopod and crinoid fragments. There is a fair to moderate amount of bioturbation. Facies A is the base of the Bakken Formation and sits unconformably on the Three Forks Formation. Facies B is composed of a silty dolostone that is gray to light brown. The facies has a high amount of bioturbation. The primary trace fossils are helminthopsis burrows. There are rare occurrences of thin laminations and brachiopod fragments. Clay content decreases near the top of the facies. Facies C is composed of a laminated silty dolostone that is light gray to light brown. Facies C is characterized by the rhythmically bedded laminations. XRD data indicates that the amount of clay content increase near the top of Facies C. The upper shale is an organic-rich mudstone that is black to dark brown in color. The shale has millimeter sized laminations, the occasional burrowed interval and contains fossil fragments. The upper contact is conformable with the Lodgepole Formation. 3.4.3 Foghorn-Ervin 20-3 (Sec. 20-T23N-R58E) The Foghorn Ervin 20-3 well is located in the south-east area of the Elm Coulee Field. The cored interval is 10,487-10,546 ft and contains the Upper Bakken Shale and five of the Bakken facies: A, B, C, E, and F (Figure A5). Facies A is composed of a silty clay-rich wackestone that is light gray to gray. This facies has abundant fossil fragments that consist of brachiopod and crinoid fragments. There is a fair amount of bioturbation and some horizontal burrows. Facies A forms the basal unit of the Bakken Formation, and sits uncomformably on top of the Three Forks Formation.

26

Facies B is composed of a silty dolostone that is gray to light brown. The facies has a high amount of bioturbation. The primary trace fossils are helminthopsis burrows. Facies B tends to have less clay content toward the top. Facies C is a laminated silty dolostone that is light gray to light brown. This facies has cross and wavy laminations and some minor bioturbation is observed. Facies E is composed of a laminated to bioturbated silty dolostone that is gray to light brown. The facies is more bioturbated at the base and has a few laminations toward the top. Facies F is a fossiliferous dolomitic to lime wackestone. The facies is dark gray to brown, and abundant brachiopod fragments are found throughout this facies. The upper shale is an organic-rich mudstone, and is black to dark brown in color. The shale has millimeter-sized laminations which become less frequent toward the top. The upper contact is conformable with the Lodgepole Formation. 3.4.3 Jackson Rowdy 3-8 (Sec. 3-T26N-R51E) The Jackson Rowdy 3-8 well is located in the north-west area of the Elm Coulee Field. The cored interval is 7,627-7,683 ft and contains both of the Bakken shales and six of the Bakken facies: A, B, C, D, E, and F (Figure A6). The lower shale is an organic rich mudstone, and is dark in color. The shale has thin millimeter sized laminations. The facies has micro-fractures that are generally cemented by calcite and pyrite. Facies A is composed of a silty clay rich wackestone that is light gray to gray. The facies has abundant fossil fragments that consist of brachiopod and crinoid fragments. There is a low to fair amount of bioturbation that increases toward the top of the facies. Facies B is composed of a silty dolostone that is gray to light brown. The facies has a high amount of bioturbation. The primary trace fossils are helminthopsis burrows. There are rare occurrences of thin laminations and brachiopod fragments.

27

Facies C is composed of a laminated silty dolostone that is light gray to light brown. The facies is characterized by the rhythmically bedded laminations. There is a low amount a bioturbation observed at the base of this section. Facies D is very fine-grained quartz-rich sandstone that is light brown. There are millimeter sized laminations found in Facies D. Facies E is composed of a laminated to bioturbated silty dolostone that is gray to light brown. The laminations are discontinuous to wavy, and become more rare toward the top of the interval where the bioturbation increases. Facies F is a fossiliferous dolomitic to lime wackestone that is dark gray to brown. Brachiopods fragments are observed in this facies, and there is a fair to moderate amount of bioturbation. The upper shale is an organic-rich mudstone, and is black to dark brown in color. The shale has millimeter sized laminations, is burrowed, and contains fossil fragments. There is some pyrite nodules observed in the upper shale. 3.4.4 Vaira 44-24 (Sec. 24-T24N-R31E) The Vaira well is located in a central part of Elm Coulee Field. The cored interval is 9,998-10,035 ft and contains the Upper Bakken Shale and three of the Bakken facies: A, B, and C (Figure A7). Facies A is composed of a silty clay-rich wackestone that is light gray to gray. The facies has abundant fossil fragments that consist of brachiopod and crinoid fragments. Bioturbation is moderate to low in this interval. Facies B is composed of a silty dolostone that is gray to light brown. The facies has a high amount of bioturbation. The primary trace fossils are helminthopsis burrows. There is a higher amount of bioturbation toward the base of the interval. Facies C is composed of a laminated silty dolostone that is light gray to light brown. The facies is characterized by the ripple and parallel laminations. There is a higher amount a bioturbation observed at the base of this section.

28

The upper shale is an organic-rich mudstone that is black to dark brown in color. The shale has millimeter sized parallel laminations. The upper contact is conformable with the Lodgepole Formation. 3.4.5 Williams 1-4 (Sec. 4-T23N-R55E) The Williams well is located in a central part of Elm Coulee Field. The cored interval is 10,047-10,072 ft and contains the Upper Bakken Shale and three of the Bakken facies: A, B, and C (Figure A8). Facies A is composed of a silt-rich wackestone. The facies is light gray to gray. The facies has abundant fossil fragments that consist of brachiopod and crinoid fragments. Bioturbation is low to absent in this interval. Facies B is composed of a silty dolostone. The facies is gray to light brown. The facies has a high amount of bioturbation. The primary trace fossils are helminthopsis burrows. There is a higher amount of bioturbation toward the base of the interval. Facies C is composed of a laminated silty dolostone. The facies is light gray to light brown. The facies is characterized by the ripple laminations. There is a higher amount a bioturbation observed at the base of this section. The upper shale is an organic rich mudstone, and is black to dark brown in color. The shale has millimeter sized parallel laminations. 3.5 Mineralogy and Diagenesis Analysis of the mineralogy and diagenesis of the Elm Coulee Field is critical component of reservoir quality. The reservoirs with the best reservoir characteristics tend to be high in dolomite, and low in calcite and clay minerals. The process of dolomitization is the key process in developing reservoir quality rock. Dolomitiztion is the exchange of calcium ion with magnesium ion in carbonate rocks. The volume of the magnesium ion is less than that of a calcite ion so the exchange can increase the pore spaces in carbonate rocks up to 13%. The importance of dolomite was noted in core studies (Figures 3.14 and 3.15). The high percentages of dolomite corresponded with high porosity and permeability values. To identify dolomite in non-cored wells,a series of cross plots were made. The plots compare the percentage of dolomite seen in the XRD data with bulk density well

29

log values. The results of this study indicate that the dolomite signature is between 2.63 g/cm3 and 2.66 g/cm3 (Figure 3.16 and 3.17). The calcite is 2.8-2.9 g/cm3 on the bulk density curves (Figure 3.17). The average bulk density in Facies B and Facies C was mapped across the study area (Figure 3.16). The trends on the map highlight that the dolomitic signature was found in southern parts of the study area. These areas tend to have the best porosity and associated production in Richland County. There are diagenetic pinch-outs of the dolomite facies to the northeast and northwest of Richland County (Figure 3.18). These pinchouts may also act as diagenetic traps that are a result of poor reservoir quality in the key facies. In addition to having better matrix porosity and permeability, the dolomite tends to be more brittle and thus more susceptible to the tectonic stress in the study area. These stresses can cause fractures in the rocks and can contribute to reservoir quality.

30

Figure 3.14. RR Lonetree, gamma ray, porosity, and permeability and XRD analysis. The well log data shows good porosity and permeability corresponding to areas with high amounts of dolomite and low amounts of calcite and chlorite.

31

Figure 3.15. Jackson Rowdy, gamma ray, porosity, and permeability and XRD analysis. The well log data shows good porosity and permeability corresponding to areas with high amounts of dolomite and low amounts of calcite and chlorite.

32

33

Figure 3.16. Brutus East Lewis cross plot of bulk density and XRD precentage of dolomite and calcite (limestone). The colored dots refer to the depth of the Middle Bakken. The red is at 7,636 ft and the dark blue is 7,678 ft. Data shows that a high precentage of dolomite corresponds to a bulk density of 2.63 g/cm3, and the calicte signature is 2.68 g/cm3 .

Figure 3.17. RR Lonetree Edna, cross plot of bulk density and XRD precentage of dolomite. The colors refer to the depth of the Middle Bakken. The red is at 10,376 ft and the dark blue is 10,408 ft. The data shows an average bulk density of 2.63 g/cm3 where there is a high precentage of dolomite.

34

35

Figure 3.18. Average bulk density map of the Middle Bakken. The average dolomite signature (2.63-2.66 g/cm3) is observed in the southern half of the study area. The calcite signature 2.68-2.69) g/cm3 is seen in the northern and western parts of the study area, and are considered as diagenetic pinch outs of reservoir properties.

CHAPTER 4

SUBSURFACE WELL LOG INTERPRETATION

Digital well logs were acquired for 421 wells in the Elm Coulee study area (Figure 4.1). All of the wells are vertical penetrations. The wells were selected on the availability of resistivity, gamma ray, and density-neutron logs. All of the digital logs were obtained from TGS Geological Products and Services. There are six wells that had conventional cores that could be used to compare with the digital logs. Wells that did not penetrate the Bakken Formation were not used in this study. Strata from the Devonian Three Forks to the Mississippian Lodgepole formations were identified and used in subsurface mapping. Using well log signatures, a series of cross sections, and structure and isopach maps were constructed. 4.1 Methods The well logs were loaded into Geographix geologic mapping and correlation software. Petrophysical log signatures of gamma ray and density-neutron porosity were compared to the six Elm Coulee whole cores to enable Bakken Formation facies correlation. The Upper Bakken Shale, Lower Bakken Shale, and Middle Bakken A, B, C, E and F facies were mapped over the Elm Coulee Field. The underlying Three Forks Formation and the Nisku (BirdBear) Formation were also mapped. Formation Tops were picked in the 421 wells and used to construct a series of northwest-southeast and northeast-southwest cross sections (Figure 4.2). Isopach and structure maps were constructed to determine the areal distribution of the individual Bakken facies. The maps also exhibit the overall geometry of the Elm Coulee Field seen in the mapping of the major units. 4.2 Structural Cross Section Eight structural cross sections were constructed across the Elm Coulee Field area, cross sections A-A ,B-B, and C-C are oriented in the northwest-southeast orientation. Cross sections D-D, E-E, F-F, G-G, and H-H are in a northeast- southwest orientation (Figures 4.3 and B1 through B7). The structural cross sections are used to

36

37 Figure 4.1. Location map of digital wells. Yellow squares mark digital well locations

contour interval is 200ft.

38
map TVDSS

Figure 4.2. Location map of structural and stratigraphic cross sections. Top Bakken structure map. Contour interval is 200ft.

determine the geometry of Elm Coulee Field. The top formation picks are used to generate depth calibrated structural surfaces that will be used in Petrel. 4.3 Stratigraphic Cross Section Eight stratigraphic cross sections were constructed across the Elm Coulee Field area, cross sections A-A, B-B, and C-C are oriented in the northwest-southeast orientation (Figures 4.4 and C1 through C8). Cross sections D-D, E-E, F-F, G-G, and H-H are in a northeast southwest orientation. The stratigraphic cross sections helped in mapping of the individual Bakken facies and their log character across the field. The thickness of the individual facies and their areal extent are displayed in a series of isopach maps. The total Middle Bakken isopach shows a thick orientation in the northwest in the middle of Richland County (Figure 4.5). 4.4 Structure Maps Structure maps were made on the top of the Upper Bakken Shale, Three Forks, and Nisku (Birdbear) formations (Figures 4.6 through 4.8) . All three fromations are present in the digital logs selected for this study. The maps show a dip to the southeast. The maps do not show any major fault features, however, there are some contour anamolies that may suggests subtle tectonic activity. These anomalies are present as bends and noses in the contours. There is also a change in dip in the north-west portion of the study area. 4.5 Isopach Maps Isopach maps were constructed for the Bakken facies and the Three Forks formations. The Bakken Formation is divided into seven facies (Figures 4.10 through 4.16). The mapping of the Bakken facies helped to determine areal extent and depositional thickness of the producing formations in the Elm Coulee Field. The top of the Bakken formation was used as a stratigraphic datum. This was used because its a conformable surface with overlying strata, and is areal extensive across the study area. The depocenter for the Bakken Formation in the study area has a northwest to southeast orientation across Richland County (Pramudito, 2008). The Bakken Formation ranges from 5.5ft to 52ft thick with an average thickness of 36ft in the study area (Figure 4.5.)

39

SW

NE

40

Figure 4.3. Structural cross sections G-G`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black).

G NE

red is facies B and C and the purple is facies A.

SW

41

light gray is facies E and F, the light

Figure 4.4. Stratigraphic cross section G-G`. The dark gray is the Upper and Lower Bakken Shales. The light gray is facies E and F, the light red is facies B and C, and the purple is facies A.

interval is 2.

5ft. The thickness ranges from 1.5ft to 5ft in the study area

42

). Contour

Figure 4.5. Thickness map of the Middle Bakken Member (facies A,B,C,D,E, and F). Contour interval is 2.5 feet. The thickness ranges from 5.5 feet to 52 feet in the study area. The Middle Bakken Member thins to the north and east and pinches out to the south and west.

The Lower Bakken facies pinches out to the south of Elm Coulee Field. The Lower Bakken Shale is only found in the northen parts of the field area (Figure 4.15). Facies A, B and C thicken in the middle of Elm Coulee, and thin to the north and pinch out to the south. The E and F faices are found restricted to the northern parts of the study area, and thicken toward the center of the basin. The Upper Bakken Shale extends over the total study area and thins toward the south. The Three Forks Formation ranges in thickness from 80ft to 180 ft. This formation thins to the southwest. Pramudito (2008) suggested that the southern portion of the study area was preferentially uplifted prior to the deposition of the Bakken Formation. The thickness anomaly in the Elm Coulee Field is a result of salt dissolution from the Devonian Prairie Salt (Sonnenberg and Pramudito, 2009). The dissolution and resulting collapse made accommodation for the Middle Bakken, and influenced the geometry of the northwest-southwest orientation of Elm Coulee. The isopach maps have been used in distributing the facies in the Petrel geomodel.

43

44 Figure 4.6. Structure Map of Upper Bakken Shale. The structure map shows a dip to the east. The contour interval is 100 ft.

45 Figure 4.7. Structure Map of the Three Forks Formation. The structure map shows a dip to the east. The contour interval is 100 ft.

46 Figure 4.8. Structure Map of the Nisku (Birdbear) Formation. The structure map shows a dip to the east. The contour interval is 100 ft.

47 Figure 4.9. Thickness map of the Upper Bakken Shale. Contour interval is one foot (0.30 meters). The thickness ranges from 4.5ft to 13ft in the study area.

48 Figure 4.10. Thickness map of the CSM F facies. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to five feet in the study area.

49 Figure 4.11. Thickness map of the CSM E facies. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to 7.5 ft in the study area.

50 Figure 4.12. Thickness map of the CSM C facies. Contour interval is two feet (0.61 meters). The thickness ranges from 2.4ft to 17ft in the study area.

51 Figure 4.13. Thickness map of the CSM B facies. Contour interval is one foot (0.30 meters). The thickness ranges from seven feet to 27 ft in the study area.

52 Figure 4.14. Thickness map of the CSM A facies. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to eight feet in the study area.

53 Figure 4.15. Thickness map of the Lower Bakken Shale. Contour interval is one foot (0.30 meters). The thickness ranges from 1.5 ft to six feet in the study area.

54 Figure 4.16. Thickness map of the Three Forks Formation. Contour interval is ten feet (three meters). The thickness ranges from 1.5ft to five feet in the study area.

CHAPTER 5

PETROPHYSICAL ANALYSIS AND RESERVOIR PROPERTIES

Petrophysics is used to determine a correlation between core properties and core identified facies with digital well log properties. This methodology is instrumental in identifying Middle Bakken rock properties and projecting them to the surrounding digital well logs. Identifying porosity, permeability, and oil and water saturations will help in the extrapolation of data allowing for a field scale study to be conducted. The data used in this study are digital well log curves: gamma ray, resistivity, neutron and density porosity, core analysis reports, and drill stem tests. There are 421 digital well logs in Elm Coulee Field that were provided by TGS geological products and services. There are six wells with routine core analysis reports. These reports have porosity and permeability and residual saturations (water,oil). The four Enerplus wells have petrographic thin sections and XRD data through the Bakken interval. The wells used in this study were interpreted using Geographix Prism log interpretation software. The log data was evaluated to generate saturation, porosity and permeability calculations which aided in reservoir characterization, and was imported into Petrel geologic modeling software for matrix modeling purposes. 5.1 Core Properties Analysis The core analysis run on the six cores provided porosity, permeability, water saturations and oil saturation data to help populate the Petrel geo-model (Figures 5.15.9). Porosity is defined as the percent of void space found in a void of rock. To evaluate effective porosity Boyles law and helium gas are used to identify interconnected pore space (Monicard, 1980). Permeability the ability, or measurement of a rocks ability to transmit fluids is typically measured in darcies or millidarcies. The term was invented by Darcy, who showed that the common mathematics of heat transfer could be modified to adequately describe fluid flow in porous media (Varhaug & Gillis, 2011). Absolute permeability is the ability to flow a single fluid through a rock. Permeability was measured by moving air

55

through a sample at steady state conditions. Fractures in core sample should be avoided because they will skew the results, and a proper matrix permeability calculation is unobtainable. Saturation is the fraction of oil, water, and gas found in a given pore space. This is expressed as a volume/volume percent of saturation units (Varhaug & Gillis, 2011). Typical saturation analysis does not show 100% fluid saturations due to the volume expansion and fluid loss associated with bringing a subsurface core with typical higher temperatures and pressures to the surface with lower temperatures and pressures. Core analysis reports from six cores within the Elm Coulee Field were analyzed: Jackson Rowdy 3-8,Brutus East Lewis 3-4-H, Foghorn-Ervin 20-3-HILD3, RR LonetreeEdna 1-13, Vaira 44-24, and Williams 1-4. Core porosity and core permeability semi-log plots were made to determine a relationship between the two variables. A logarithmic regression line was then added to fit the data. A correlation coefficient number R 2 was then used to compare the plots to the regression line to determine if a relationship could be determined. This was a good exercise that showed the difference between dolomite rich, relatively higher porosity and permeability reservoir rocks (Foghorn and Lonetree)(Figure 5.1) from limestone and clay rich rocks that have relatively lower porosity and permeability (Jackson Rowdy and Brutus)(Figure 5.1). The results from the total Middle Bakken interval using all well data points are plotted in Figure 5.2. 5.2 Conventional Well Log Analysis Well logging is procedure of measuring depth or time, or both, of one or more physical quantities in or around a well bore (Varhaug & Gillis, 2011). These geophysical measurements are used to identify the electrical properties, porosity, lithology, mineralogy, permeability and saturation of the subsurface environment (Asquith et al, 2004). The gamma ray log investigates the total natural radioactivity, measured in API units. The measurement can be made in both openhole and through casing. The depth

56

57 Figure 5.1 Porosity and Permeability cross plots. The Foghorn and Lonetree Wells have good logarithmic regression line correlation. The Brutus and the Jackson Rowdy well have poor logarithmic line colorations. The former two wells are situated in a dolomitic setting as indicated by bulk density and XRD analysis.

Core Data
10.0 9.0 8.0 7.0 Porosity % 6.0 5.0 4.0 3.0 2.0 1.0 0.0 0.0001 0.0010 0.0100 0.1000 Permeability (md) Core Data Log. (Core Data) y = 1.1091ln(x) + 11.804 R = 0.5277

Figure 5.2. Middle Bakken porosity and permeability crossplot (from Enerplus core data)

of investigation is a few inches, so that the log normally measures the flushed zone (Varhaug & Gillis, 2011). The neutron porosity log measures formation porosity by measuring the fast neutrons emitted by a source. Hydrogen has the biggest effect on the slowing and capturing of neutrons. Hydrogen is primarily found in pore fluids. The matrix and type of fluid have an influence on the log measurements. The logs used in this study were calibrated to limestone matrix. The density porosity log is used to measure formation density. This is accomplished by monitoring the flux of gamma rays between a source and a detector due to Compton scattering. The bulk density is calculated using the following equation,

b * f (1 )ma
where b is bulk density, is porosity, ma is matrix density and

f is fluid density. For

this study a matrix density of 2.71 gm/cm3 was used and value of 1 was used for fluid

density.

The porosity from well logs was calibrated to cored data by using an average from the density porosity and neutron porosity. This is achieved by applying the equation

58

n 2 d 2
2

where n and d are neutron and density porosity respectively. The combination of these log properties is largely free of lithology effects and tends to correspond to core

porosity values better than the individual neutron and density porosity logs (Figures 5.4 through 5.9). The average porosity was mapped across the study area (Figure 5.10) and has a positive correlation to the high dolomite percentages seen in XRD analysis (Figures 3.16 and 3.17). Electric logs can either measure the resistivity or conductivity of a formation. The pore fluids can influence this measurement by their ability to conduct the current flow from the tool. A permeability curve was calculated using the relationship observed in the Middle Bakken porosity permeability cross plot (Figure 5.2) from the Enerplus cored wells. This was achieved by applying the equation:

exp(
where

11.804
1.109

is the average porosity calculated from neutron and bulk density logs. The

correlation between calculated permeability and core permeability can be seen in figures 5.10 and 5.11 The water saturation (Sw) was generated using parameters from Holmes (2005). Holmes calculated a formation water resistivity (Rw) of 0.03. This allows for the construction of a hydrocarbon saturation, porosity and net pay thickness map (SoPhiH) (Figure.5.12).

Sw n

Rw m Rt

Where n is 1.74, m is the cementation component and equal to two, Rt is the true formation resistivity.

59

B .

Figure 5.3. Photomicographs showing examples of nonreservoir (A) and good reservoir (B) in Facies B. Florescent light is used to highlight porosity in dyed thin sections. The orange color indicates porosity. (from Alexandre, 2011)

60

Figure 5.4. Foghorn Ervin Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core So), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix.

61

Figure 5.5. Brutus East-Lewis Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core So), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix.

62

Figure 5.6. Lonetree-Edna Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core So), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix.

63

Figure 5.7. Jackson Rowdy Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core So), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix.

64

Figure 5.8. Vaira Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core So), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix.

65

Figure 5.9. Williams Core Data: gamma ray (GR), bulk density (RHOB), neutron porosity (NPHI), water saturation (Core Sw), oil saturation (Core So), porosity, permeability (Core Kmax). Log templates constructed in Prism module of Geographix.

66

average porosity map of Elm Coulee Field.

67 C and Facies B

Figure 5.10. Facies C and B average porosity map of Elm Coulee Field. The best porosity is found in the southern half of the study area.

Figure 5.11. Jackson Rowdy, thick red lines in the Porosity and PERM tracks are the calculated average porosity (PHIA) and permeability (Perm).

68

Figure 5.12. Foghorn Ervin, thick red lines in the Porosity and PERM tracks are the calculated average porosity (PHIA) and permeability (Perm).

69

70 Figure 5.13. Middle Bakken SoPhiH Map indicating the most petrophysically prospective areas of Elm Coulee Field.

5.3 Drill Stem Tests Drill Stem testing (DST) is a procedure to determine the productive capacity, pressure, permeability or extent (or a combination of these) of a hydrocarbon reservoir. The purpose is to isolate the zone of interest with packers. Next, one or more valves are opened to produce the reservoir fluids through the drill pipe, and allow the well to flow for a time. Depending on the requirements and goals for the test, it may be of short (one hour or less) or long (several days) duration and there might be more than one flow period and pressure buildup period. Horner Plots were constructed using DST data to determine reservoir pressure and calculate permeability found in Elm Coulee. Plots were built for four wells, Parsons 14-27, Sorenson 8-17, Vaira 44-24, and Bahl 12-1 (Figure 5.13). The Sorenson 8-17 well is in the same section as the Foghorn well, and the Bahl well is located in the same section as the Lonetree well. Permeability was calculated from Horner plots by fitting a linear regression to the steady state portion of the built up curve, the calulated slope is then used as a measure of permeability. This made it possible to compare core data with DST data. The DST data was compared with average core data in the study area, a summary of results is listed in the Table 2. Comparing permeability from core and the permeability from the DSTs reveals that there is, on average, an order of magnitude of difference between the two values. This may suggest that fractures are enhancing reservoir properties beyond what is seen in core data. The calculated formation pressures indicate that the field area is over-pressured; this may also contribute to production in the field area. The overall results from DSTs are poor and should be considered inconclusive, because the DSTs were unreliable from a fluid sampling point of view.

71

Figure 5.14. Horner plots of Elm Coulee Wells

Psi/ft

md

md

Table 2. Summary of service company pressure and DST results, compared to average core permeability data (The Sorenson well is in close proximity to the Foghorn well, and the Bahl Well is located close to the Lonetree well).

72

CHAPTER 6

FRACTURES
Fractures, either induced or natural, play an integral role in developing resource plays. Unconventional plays are known for their characteristically low porosity and permeability, and natural fractures can contribute to reservoir porosity and permeability. An understanding of stress directions and natural fracture patterns within the basin are imperative in well design and completion. By exploiting natural fractures with an optimum well design, an operator can ensure maximum hydrocarbon recovery. This purpose of this chapter is to describe a fracture model for the Elm Coulee Field. The fracture model is used to construct a dual porosity model in Petrel that incorporates both matrix and fracture properties. Using regional fracture trends and available field data a fracture model in the Elm Coulee Field has been developed. 6.1 Production Data. Production data was acquired from IHS Production Energy Information. The data was used to plot cumulative production, first year cumulative production, and to calculate an exponential decline curve to derive estimated ultimate recovery (EUR) for the producing wells in Elm Coulee. The wells used in the production calculations were selected by filtering for wells with horizontal well bores. The most crucial element of the production data that helped identify fractures was the first year cumulative data. Plotting production data helped to identify sweet spots in Elm Coulee Field. Sweet spots are areas of high production and show a steady distribution throughout the Elm Coulee field (Figure 6.1). The sweet spots have an ovoid shape that roughly trends north and south. The sweet spot production can be influenced by operator bias which involves fracking and completion techniques that either maximize production or minimize production to ensure optimal production for the life of the well. The operators were mapped across the study area and displayed over the first year cumulative map and there was no clear indication of operator bias influencing the production results (Figure 6.2). Sweet spots are hypothesized to be caused by a combination of both matrix properties and fracture properties. The intersection of these two properties is an area that correlates to the best production trends in the study area. To get a better idea of the fracture trends the first year cumulative production contours were introduced to a bias 73

trend that is similar to the maximum principal stress of the Williston basin. The bias was oriented N60E and resulting product of this exercise are a series of elongate contours (Figure 6.3). The production helped to interpret possible regional fracture swarms that have a similar strike and location to anomalies in the top Bakken structure map, and in the Middle Bakken isopach map (Figures 4.5 and 4.6). The contours in the structure maps show nosing and strike rotations in the sweet spot trends. The isopach map shows left lateral offsets in the northeast and southwest which is perpendicular to the northwest and southeast deposition trend. The fracture trends are spaced approximately 4.75 miles apart and have an average contour length of 5.7 miles. These production trends are found in the southern portions of the study area, an area that has a strong dolomite signature and would be more susceptible to brittle failure. 6.2 Other Available Data A micro seismic study by OBrien, Larson, and Parham ( 2011) shows that the artificial induced fractures are consistent with the maximum principal stress directions of the Williston basin. Seismic attribute mapping done by Angster, (2010) highlights some of the regional fracture spacing and lengths seen in the Bakken Formation. Using these available data helped construct a conceptual fracture for Elm Coulee Field. 6.3 Seismic Mapping in Billing Nose an analog for a local fracture fabric of the Elm Coulee Field. A previous study at the Colorado School of Mines by Angster (2010) interpreted a 3D seismic data volume from the Billing Nose area of western North Dakota. The data covered approximately 117 square kilometers of the Bicentennial Oil Field. Angster Identified fractures in the Bakken horizon using the minimum and maximum curvature attribute (Figure 6.4). The results of his study showed a dominant fracture orientation to the northeast and an orthogonal set to the northwest. The fractures have an average spacing of 1500 ft, and an average length of 2600 ft. The orthogonal set of fractures were observed to have approximately half the intensity, greater spacing, and comparable lengths as the primarily fracture trends.

74

Figure 6.1. First year cumulative production

Figure 6.2. First year cumulative production and Operators

75

76 Figure 6.3. Initial production trends with a N60E Bias. This bias data trend highlights the fracture swarms that are interpreted in the study area

77 Figure 6.4 Minimum curvature attribute map, Bakken horizon ( from Angster 2010). Interpreted fractures are on the right map. The red colors show the northeast orientation and the yellow colors show the northwest orientation. The fractures in this study have a similar orientation to stress directions found in Williston basin.

6.4 Micro Seismic Study Elm Coulee Field. In 2011, a study of the artificially induced fractures from a multi-stage fracking process was documented in a paper by (O'Brien, Larson, & Parham, 2011). The purpose of their study was to determine fracture azimuth, fracture height and half length. The study was located north of the Vaira well in a central area of Elm Coulee Field (Figures 6.5 and 6.6). The authors reported an average fracture azimuth of N63E, fracture height of 290ft and half-length of approximately 450ft. The spacing between the artificial induced fractures is on the order of 1200 ft. The spacing observed in the artificial fractures is similar to the fractures spacing seen in Angsters (2010) study.

Figure 6.5 Summary of micro seismic observations. Average fracture azimuth N63E. Fracture half lengths averaged 450 ft (from O'Brien, Larson, & Parham, 2011).

78

Figure 6.6 Micro seismic production traces and fracture height data. The average fracture height was 290 ft. Radioactive tracer show zonal isolation averaged approximately 1200ft (O'Brien, Larson, & Parham, 2011). 6.5 Results The results of these fracture studies allowed for the formulation of a conceptual fracture model. The model will incorporate fracture swarm trends, and a regional fracture trend. The fracture swarms are interpreted to be controlled by underlying basement trends that accommodate large scale fracture fairways identified from initial production. The fairways have a fracture spacing of 4.75 miles and an average fracture length of 5.7 miles. The regional fractures are considered a fracture fabric associated with the Elm Coulee Field area. This was derived from the seismic attribute mapping done by Angster and reinforced by the work done by OBrien, Larson and Parham, (2011). These two studies indicate a maximum principal stress direction to the northeast, and natural fracture spacing on the order of 1200-1500 ft. The work done by Angster (2010) highlights the orthogonal set of fractures in the northwest direction, with about half the intensity as those of the maximum principal stress direction.

79

A summary of the data is displayed in Figure 6.7. The figure shows the fracture density per mile plotted against the fracture lengths observed in the studies. A power log relationship is used to correlate the studied fracture patterns with a regression line. The data was compiled to construct a fracture model that incorporates all of the data sets. It is interpreted that the fractures fairway derived from production data have the greatest influence on Elm Coulee production (Figure 6.8). These regional trends would have zones of deformation that would influence the regional fracture fabric and facilitate increased production. The regional fracture fabrics determined from microseismic and 3D seismic are interpreted as having an influence on artificially induced fractures, but are not interpreted as major contributors to sweet spot production. They may contribute locally to enhanced reservoir properties as observed in the difference between core data and DST data. The fracture swarms are interpreted to have the greatest influence on fracture fabric and contribution to production. The swarms are interpreted as the results of tectonically controlled basement faults with subtle offsets. The behavior was observed in Angster (2010).

80

81

Figure 6.7. Fracture lengths versus fracture spacing. The logarithmic plot shows the power log relationship between the production, seismic and micoseismic data used in the fracture model. The data shows outliers and these may be the result of having uncertainty in isolating the fracking stages in the micro-seismic, and the use of seismic data located east to of study area but not in Elm Coulee, However, the overall data does have a good correlation R2 of 82 percent

82 Figure 6.8 Conceptual fracture model. The fracture swarm trends in the blue ovals and are spaced ~4.75 miles or 25,000ft apart. The swarms have the greatest influence on the fracture fabric. The green lines reflect the regional fracture fabric that are oriented with the maximum principal stress and have a spacing of ~1,250ft as was seen in both the microseismic and Bicentennial seismic survey. The orange lines show the orthogonal spacing of fractures that are spaced ~2,500ft apart.

CHAPTER 7

3-D GEOLOGICAL MODELING

Developing a 3-D geologic model captures Elm Coulee reservoir properties for the purpose of running reservoir simulation, and for computing numerical petroleum relationships. The reservoir properties of porosity and permeability must be calibrated so that they can match historical production of the wells in the study area. The geologic framework is incorporated in the model before any simulation can be run. The geologic and seismic data, and how they were interpreted have been explained in previous chapters. These data and interpretations were used to build a geomodel in an effort to match production results. 7.1 Modeling in Petrel The 3-D model constructed in Elm Coulee Filed was completed using the Schlumberger Petrel software version 2010.1. The support of the regional Schlumberger office in Greenwood Village, Colorado greatly aided in building the geomodel. The first step of the modeling was to load all available data into the program. The data include well names, unique well identification number (UWI), well location, reference elevations, digital well logs, formation tops and bases. The digital wells used to build the model were all vertical, and therefore did not require directional surveys. The reservoir model was completed using a data subset of the total field area (Figure 7.1). The reason for this was to reduce the size of the model to help with computational processing speed. The area selected incorporated two cored wells, and has features displaying interpreted matrix and fracture properties. Using formation and facies well tops of the Lodgepole, Bakken, and Three Forks formations, structure maps were built in Geographix petroleum geology software. The contoured surfaces were converted into grid layers that were used as three dimensional surfaces and imported into the Petrel static model. Once the structure and thickness of the model are imported, the model is populated with cells forming a geo-cellular grid (Figure 7.2). The cells are then populated with lithological, and reservoir properties derived from core and petrophysical analyses.

83

Figure 7.1. The red box is the outline of the Petrel model and shows the location of the digital wells used to distribute petrophysical properties.

The grid cell size is chosen for its ability to show important geologic properties and also allows a reasonable time for computer simulation runs. The gird size chosen for this project is 100x100 ft. This gird cell size allowed for the capture of hydraulic fracture spacing intervals that can be modeled in a producing Bakken well. The geo-cellular grid was oriented in the maximum principal stress direction N60E, to best capture the fractures in the study area which are perpendicular to the majority of the drilled horizontal wells.

84

85

Figure 7.2 Petrel geo-model structural model and geo-cellar grid. The mapped horizons are the Three Forks, Bakken, and Lodgepole formations.

Data from the well logs and core are used to distribute these petrophysical properties(Figures 7.3 and 7.4). The petrophysical properties used were the calculated average porosity (PHIA) and the derived permeability (Perm) from the core data cross plot (Figure 5.10 and 5.11). The permeability and porosity were converted into digital well logs that could be imported into Petrel software. The wells used (Appendix D) were selected on their availability of digital logs, and the ability to calculate reservoir properties. Wells needed to have resistivity, porosity and gamma ray well log curves. To distribute petrophysical properties between well data control points a geostatistical method uses an algorithm to populate properties between cells. There are a number of algorithms, and a statistical control to execute this property modeling that can vary the results.

Figure 7.3. Middle Bakken porosity distribution 3D view looking north. Purple is low (0.0%) and red is high (20.0%)

86

Figure 7.4. Middle Bakken permeability distribution. 3D view looking north. Purple is low (0.001mD) and red is high (0.1mD) log scale

Geologic observations used to develop model parameters include the basin geometry, facies and lithological variations, and petrophysical. A deterministic modeling method was due used to construct the model. The density of the well logs and the ability to upscale the well-log data was a primary reason for using this method. This type of modeling is faster to run and gives a single estimated result. The averaging method used for porosity was an arithmetic mean which is recommended for this attribute. The preferred averaging method for the permeability was a harmonic mean averaging which works well for attributes with log normal distribution. The algorithm for distributing all the properties is a Gaussian Random Function Simulation developed by Petrel. This method was used to interpolate between wells. The goal of this study was to create a conceptual model that could be used in reservoir simulations, since this phase of the study did not run a reservoir model multiple realizations were not produced ,however, this should be considered in future work to ensure all avenues are perused.

87

7.3 Fracture Properties The fractures are an attribute that may contribute to production in the study area. There are two fracture features that are considered. There are the regional fracture patterns seen in seismic and micro-seismic, and the fracture swarms observed from production data that is interpreted to be associated with basement faulting. The regional fracture fabric is oriented in two directions. The major fracture set is oriented to the northeast in the maximum principal stress direction N60E and the orthogonal fracture orientation is N150E. The fractures are spaced according to the grid size used in the model and reflect the spacing outlined in Chapter 6 (Figure 7.5).

Figure 7.5 Localized fractures derived from Centennial seismic and the micro-seismic study from Enerplus. The dominant fracture trend is the northeast orientations (blue polygons) and the orthogonal northwest fracture set is represented by the purple polygons.

The grid area is 100 by 100 ft. The ratio of 0.08 fractures per unit area (F/A) was used for the principal stress direction, and 0.04 F/A was used for the orthogonal fracture

88

orientation. This ratio resulted in a fracture being placed every 1,250 ft in the maximum principal stress direction, and every 2,500 ft in the perpendicular orthogonal direction. The next step was to relate the localized fracture pattern to the initial production trends allowing for the most intense fracture contributions to come from the identified fracture swarm fairways (Figure 7.6). To accomplish this, first year cumulative production data was contoured and made into a surface, and then normalized. Localized fracture models were made for the two fracture directions displaying both observed spacing and lengths. The two localized fracture models were then multiplied by the normalized surface. The normalized surface had a high of 0.08 F/A in the northeast direction, and high of 0.04 F/A in the northwest direction respecting the localized fracture intensity patterns.

Figure 7.6. The figure shows the relationship of the two fracture patterns 1. regional fracture pattern (purple lines) and 2. the contoured first year cumulative production data, cropped from figure 6.3. The contoured interval is 2,000 barrels of oil. The orange is 18,000 barrels and the blue is 4,000 barrels of oil.

89

The final result is a fracture model that has both the maximum principal stress direction fractures, and the orthogonal fractures that varied in fracture intensity according to the regional fracture model (Figure7.7 and 7.8). The model is an interpreted fracture model that is related to production at Elm Coulee. The fracture model and the matrix model were combined to construct a dual porosity model (Figure 7.9).

Figure 7.7 Localized fractures derived from Centennial seismic and the mico-seismic study from Enerplus. The fractures have been related to the first year cumulative production data and have varying degrees of fracture influence.

90

Figure 7.8. Oblique view highlighting the fracture variability in the fracture model. The blue polygons are the northeast fractures, and the red polygons area the northwest fractures.

Figure 7.9. Dual porosity model. Model shows the matrix model combined with the fracture model creating a dual porosity model.

91

CHAPTER 8 DISSCUSSION

The purpose of this research was to: (1) to complete an in-depth examination of the reservoir properties of the Bakken reservoir system in Elm Coulee field; (2) to construct a three-dimensional geologic model that shows the distribution of the different facies within the Bakken Formation, and their associated reservoir properties: and (3) to build a fracture model, and integrate it with the matrix porosity model. The three-dimensional modeling in Petrel allowed for the integration and assimilation of all the primary goals, and was achieved by integrating all the available data and interpretations. The model is limited by the quality of the data available and its geologic interpretation. The work flow used to establish the final geomodel is illustrated in (Figure 8.1). The data brought into Petrel was used to develop four different models, structuralstratigraphic, matrix porosity, matrix permeability, and fracture. The models described in this chapter where built to use as an integrated geomodel that could be used in reservoir fluid flow simulations. The model is built to be tested using both a matrix dominated reservoir interpretation, and a dual matrix and fracture reservoir interpretation. This gives flexibility when trying to compare and contrast the effects of natural fractures on production. 8.1 Structural-Stratigraphic Model The structural model was constructed using the formation and facies tops that were identified in well logs. The resulting contoured surface revealed a relatively flat stratigraphy that dipped toward the central Williston basin. The structural model also revealed subtle contour anomalies, these anomalies correspond with the fracture swarm fairways interpreted from first year cumulative production data. The structural model shows the pinch out of the Middle Bakken toward the south, which provides the stratigraphic trap to the south of Elm Coulee. The Petrel model was built in a central location of Elm Coulee, and only shows the pinch-out of the E and F lithofaces. Lithofacies were mapped according to the core descriptions and the digital log signature relationship. There is a high amount of confidence in distribution of the 92

lithofacies throughout the study area. The model was able to incorporate and distinguish all the Middle Bakken lithofaces. Due to the thin nature of the E and F facies these two facies were combined.

Figure 8.1 Diagram of the integrated geomodeling workflow showing the different component models 8.2 Porosity Model The porosity model was developed using the correlation from core data to digital log data. This resulted in calculating an average porosity using density and neutron porosity that corresponded to the core porosity values. The porosity was shown to have a high correlation with the lithologies that contained high amounts of dolomite. The porosity trend was oriented to the northwest on the southern portion of Richland County (Figure 5.10). 8.3 Permeability Model The permeability model was developed from the porosity permeability cross plot (Figure 5.3). The cross plot show a positive correlation between the two attributes. The

93

data derived from the cross plot showed the same amount of variability and trends. This data was used to construct permeability logs (Figure 5.11 and 5.12) that were used to distribute permeability in the study area (Figure 7.3) 8.4 Fractures Model The fracture model was developed using two scales of investigation, a regional fracture fabric trend, and a regional fracture swarm trend. The two interpretations were then combined to build a fracture model for Elm Coulee (Figure 6.8). The regional fracture swarm trend was developed using sweet spots in initial production in Elm Coulee, and were then trended to reflect the maximum principal stress direction in the Williston basin (Figure 6.3). The potential for operator bias influencing sweet spots of increased production is a concern. Operators use a variety of hydraulic fracturing and completion techniques that can influence the production in wells.

Figure 8.2 Initial first year cumulative production vs. Operator. The data does not show any definite production trends associated with individual operator. Production versus operators data used in Elm Coulee was plotted, and is shown in figure 6.2 and 8.2. No operator bias has been identified.

94

The fractures that are observed in core tend to be filled with calcite cements which may be the case in areas of the field without the influence of regional fracture fairways. The regional fracture fairways are interpreted as clusters of fractures that have a greater probability of hosting open mode fractures whose aperture will contribute to permeability in the surrounding Middle Bakken reservoir. The fracture fabric trend was developed using seismic and microseismic results (Angster, 2010) (O'Brien, Larson, & Parham, 2011). These fracture trends reflected the maximum principal stress directions of the Williston basin. The results of this study show fracture spacing on the order of 1,250 ft, and a fracture length of approximately 2,400 ft (Figure 7.5) . The two fracture trends were combined to highlight the potential fracture contribution in Richland County. The resulting figures 7.6 through 7.8 show increased fracture intensity oriented with the fracture fairways identified from first year cumulative production. 8.6 Production Data Production data was used to analyze the validity of the dual porosity model hypothesis (Figure 8.3). The well production was analyzed using a hyperbolic decline to calculate estimated ultimate recovery (EUR)(Arps, 1956). Hyperbolic decline is known to overestimate the total production due the assumption that reservoir conditions governing the flow rate remain constant over changes in time and pressure (Stotts, 2007). With this being said, the EUR calculations primary purpose is to provide a standard measurement in which the different trends in the study could be compared. The data suggested that there is significant EUR difference between wells with good matrix properties and wells that benefited from both good matrix properties, and the interpreted fracture fairways (Figure 8.4). Better production occurred in areas where there was greater than six percent matrix porosity, and the interpreted fracture trends overlapped. Matrix trends occurred in areas where there was greater than six percent porosity, (Figure 5.10) but did not include areas with the overlapping fracture trends. The Elm Coulee field production data was also analyzed for areas with less than six percent porosity, and incorporated all the wells in the study area. The plotted EUR data shows that the large accumulations are primarily found in the interpreted fracture fairways (Figure 8.5) 95

Figure 8.3. EUR well data. The graph shows the EUR production trends (>6%porosity+ fractures, >6%porosity-fractures,<6%porosity. The P90 and P10 were calculated by fitting a regression line to the data trends and finding the intercept that corresponded to 1.28 standard distribution.

Figure 8.4. EUR well data. The table shows the difference between matrix and enhanced (matrix and fracture trends) production. The mean production shows average well production for each well in the trends.

96

Figure 8.5. Fracture trends and a bubble map with calculated EUR Production.

97

CHAPTER 9 CONCLUSIONS

9.1 Conclusions The study has constructed separate component models for the Bakken Formation in Elm Coulee. The models include: A structural model/facies model, porosity model, permeability model, and fracture model. The Petrel geocellar gird is orientated parallel to the maximum principal stress direction. This is to take advantage of the preferred hydraulic fracturing directions in artificially stimulated wells. The following conclusion were made from the models described.

1) The porosity and permeability studies show a positive correlation in the cross plot (Figure 5.3) that allowed for a permeability model to be constructed from reservoir porosity properties. The reservoir porosity and permeability parameters seen in DSTs do not reflect those seen in average matrix properties, suggesting a fracture contribution to production. (Table 2).

2) X ray diffraction analysis studies show that the best reservoir properties are associated with lithologys that have a high amount of dolomite. The dolomite distribution in the Middle Bakken is shown to reside in the southern portion of Richland County.

3) Production is shown to correlate to areas with the best matrix properties, but the production is not uniform throughout these matrix trends. When observing initial production rates, enhanced production areas were identified. These areas are interpreted to be a combination of matrix and fracture properties.

4) Regional fracture fairways are identified from enhanced production. The fractures are modeled to reflect the regional stress directions. The local fracture patterns and spacing is derived from seismic and micro-seismic studies. The localized fracture model is independent of lithofacies and is interpreted to contribute to production when it corresponds with the fracture swarm trends.

98

9.2 Recommendations The model was developed to construct a geological template that would allow for the development of a reservoir simulation. The model needs to be tested in a reservoir simulation program. A history match is needed using the matrix model and then the dual porosity model (matrix and fractures). If there are favorable results confirming a good geo-model, then a second simulation could be performed looking at using enhanced recovery methods such as water and gas in Elm Coulee Field. The study bases its fracture interpretation on production trends and seismic data that was acquired in a nearby field. To validate the results it is recommended that a similar attribute study to Angster (2010) be conducted on 3D seismic data acquired in Elm Coulee Field. The study used logs and core data to characterize reservoir properties. The core data was sampled using plugs situated in approximately 1 foot increments. This was an effective method to correlate digital logs to core data. The potential issue with this method is that it does not capture the fracture contribution to reservoir properties. A better model may be attained by using whole core data and running modern reservoir tests, such as production logs, and extended drill stem tests. Dolomite was seen as an important factor in developing reservoir quality and could be a major attribute in the formation of fractures. The mechanisms responsible for dolomitization in Elm Coulee are not explicitly known and a geochemical study might help determine the processes involved in creating the best reservoir properties in Elm Coulee.

99

REFERENCES CITED

Alexandre, C., 2011. Reservoir Characterization and Petrology of the Bakken Formation, Elm Coulee Field, Richland MT. Colorado School of Mines, Department of Geology. Golden: Colorado School of Mines.175 p. Angster, S., 2010. Fracture Analysis of the Bakken Formation, Williston Basin. Field Studies in the Little Rocky Mountains and Big Snowy Mountains, MT and Beartooth Mountains, WY, and 3D Seismic Data, Williston Basin. Colordo School of Mines, Department of Geology. Golden: Colordo School of Mines.67 p. Arps, J.J., 1956. Estimation of Primary Oil Reserves Reserves. Trans. AIME, Vol 207, 182-191p. Asquith, G., Krygowski, D., Henderson, S., & Hurley, N., 2004. Basic Well log analysis (2nd ed., Vol. AAPG Methods in Exploration Series 16). AAPG. 240 p. Blakey, R., 2005. North American paleogeographic maps, early Mississippian (360Ma). Retrieved June 15, 2011, from http://jan.ucc.nau.edu/rcb7/namD360.jpg Brown, D., & Brown, D. ,1987. Wrench-Style Deformation and Paleostructural Influence on Sedimentation In and Around a Cratonic Basin. Rocky Mountain Association of Geologist. 57-70 p. Canter, L., Skinner, O., & Sonnenfeld, M. ,2009, January 27,. Facies and Mechanical Stratigraphy of the Middle Bakken, Mountrail County,North Dakota. Rocky Mountain Section of SEPM Luncheon. Coskey, R., & Leonard, J. ,2009, June 7-9. Bakken Oil Accumulations- What's the Trap? Presentation Session, Search and Discovery Article #90090. Denver: American Association of Petroleum Geologists Annual Convention,. Crowley, K., Judson, A., & Naeser, C. ,1985. Origin and epeirogenic history of the Williston Basin: Evidence from fission-track analysis of apatite. Geological Society of America, 620-623 p.

100

Gale, J. ,2002. Specifying lengths of Horizontal Wells in Fracured Reservoirs. SPE Reservoir Evaluation and Engineering, 266-272 p. Gerhard, L., Anderson, S., & Fischer, D. ,1990,. Petroleum geology of the Williston Basin (Vol. Interior cratonic basins: AAPG Memoir 5). (M. Leighton, D. Kolata, D. Oltz, & J. Eidel, Eds.) AAPG. 507-559 p. Green, A., & Weber, W. ,1985. Evolution of Proterozoic terrains beneath the Williston Basin. Geology 13, 624-628 p. Holmes, M. ,2005. Petrophysical analysis of the Bakken interval, Nance Petroleum, Larson 11-26 well, Richland County, Montana: Bakken Play Essentials, Short Course #1. AAPG Rocky Mountain Section Meeting course notes, 45-54 p. Meissner, F. ,1978. Petroleum geology of the Bakken Formation, Williston basin, North Dakota and Montana,. In D. Rehrig, The economic geology of the Williston basin: Proceedings of the Montana Geological Society, 24th Annual Conference , 207227 p. Monicard, R. ,1980. Properties of reservoir rocks: core analysis. Houston: Gulf Publishing Company, 168 p. Nelson, J. ,1992. Basment Control of Recurrent Faulting, Central Montana. Tenth International Conference on Basement Tectonics. Duluth, Minnesota: Kluwer Academic Publishers. O'Brien, D., Larson, R., & Parham, R. ,2011. Using Real-Time Downhole Microseismic to Evaluate Fracute Geometry for Horizontal Packer Sleeve Compleations in the Bakken Formation, Elm Coulee Field, Montana. SPE International, 22 p. Pitman, J., Price, L., & Lefever, J. ,2001. Diagenesis and gracture development in Bakken Formation, Williston Basin: Implications for reservoir quality in the middle membger. U.S. Geological Survey Professional Paper 1653, p. 17 p. Pramudito, A. ,2008. Depositional facies, diagenesis, and petrophysical analysis of the Bakken Formation, Elm Coulee field,Williston Basin, Montana. Master of Science thesis. Colorado School of Mines.

101

Price, L., Ging, T., Daws, T., Love, A., Pawlewicz, M., & Anders, D. ,1985. Organic metamorphism in the Mississippian - Devonian Bakken shale, North Dakota portion of the Williston Basin. In J. Woodward, F. Meissner, & J. Clayton, Hydrocarbon Source Rocks of the Greater Rocky Mountain Region (pp. 83-134). Denver, Colorado: Rocky Mountain Association of Geologists. Stotts, W.J., Anderson, D.M., & Mattar,L., 2007. Evaluating and Developing Tight Gas Reserves-Vest Practices. SPE Rocky Mountain Oil and Gas Symposium, Denver,CO, SPE Paper #108183. Simenson, A. ,2010. Depositional Facies and Petrophysical Analysis of the Bakken Formatoin, Parshall Field, Mountrail County, North Dakota. Golden: Colorado School of Mines, 198 p. Smith, M., & Bustin, R. ,2000, July. Late Devonian and Early Mississippian Bakken adn Exshaw Black Shale Source Rocks, Western Canada Sedimentary Basin: A Sequence Stratigraphic Interpretation. AAPG Bulletin, 84(7), 940-960 p. Sonnenberg, S., & Pramudito, A. ,2009, September. Petroleum geology of the giant Elm Coulee Field, Williston Basin. AAPG Bulletin, 93(9), 1127-1153 p. Sutter, J. ,2006. Facies Model Revisited: Clastic Shelves in Posamentier. Society for Sedimentary Geology (SEPM), 375 p. Varhaug, M., & Gillis, G. ,2011. Retrieved 2011, from Schlumberger Oilfield Glossary: http://www.glossary.oilfield.slb.com Vigrass, L. ,2001, September 1. Depositional framework of the Winnipeg Formation in Manitoba and eastern Saskatchewan. The Geological Associsations of Canada, Special Paper, 225-234 p. Walker, B., Powell, D., Rollins, D., & Shaffer, R. ,2006, June 11-13. Elm Coulee Field Middle Bakken member (Lower Mississippian/Upper Devonian) Richland County, Montana,. Billings: American Association of Petroleum Geologists Rocky Mountain Section. 375 p. Walker, R. ,2006. Facies Models Revisited in Posamentier, H. Seciety for Sedimentary Geology (SEPM), 375 p. 102

Webster, R. ,1984, Petroleum Source Rocks and Stratigraphy of the Bakken Formation in North Dakota. (J. Woodward, F. Meissner, & J. Clayton, Eds.) Hydrocarbon source rocks of the greater Rocky Mountain region, 57-82 p. Williams, H., Hoffman, P., Lewry, J., Monger, J., & Rivers, T. ,1991. Anatomy of North America: Thematic geologic portrayals of the continent. Tectonophysics, 187, 117-134 p.

103

APPENDIX A CORE DESCRIPTIONS

Five cores were described showing lithofacies, biotrubation and sedimentary structures associated with the Bakken Formation in the Elm Coulee study area. Refer to chapter 3 for details.

Figure A1 Summery of Bakken Formation lithofaces used in figures A4-A8. (from Alexandre, 2011)

104

Figure A2. Legend of symbols used in the interpretations and measured sections of figures A4-A6. This bioturbation index was first used by Reineck (1963) (from Alexandre, 2011)

105

Figure A3. Legend of symbols used in the interpretations and measured sections of figures A7-A8 (from Pramudito, 2008)

106

Figure A4. RR Lonetree-Edna 1-13 Core description (from Alexandre, 2011)

107

Figure A5. Foghorn-Ervin 20-3-H Core description (from Alexandre, 2011)

108

Figure A6. Jackson Rowdy 3-8 Core description (from Alexandre, 2011)

109

UBS

Figure A7. Vaira 44-24 Core description (modified from Pramudito, 2008)

110

UBS

Figure A8. Williams 1-4 Core description (modified from Pramudito 2008)

111

APPENDIX B STRUCTURAL CROSS SECTIONS

A series of eight stratigraphic cross sections were made to document the structure of the Elm Coulee field and use that data to build a stratigraphic frame work in Petrel. Refer to chapter 4 for details.

112

A NW

A SE

Figure B1. Structural cross sections A-A`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). 113

B NW

B SE

Figure B2. Structural cross sections B-B`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black).

114

C NW

C SE

D SW

D NE

Figure B3. Structural cross sections C-C`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black).

Figure B4. Structural cross sections D-D`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black).

115

E SW

E NE

Figure B5. Structural cross sections E-E`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black). F SW F NE

Figure B6. Structural cross sections F-F`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black).

116

H SW

H NE

Figure B7. Structural cross sections H-H`. The Bakken (Purple), Three Forks (Green) and Birdbear (Orange) are formation tops correlated across the study area. The track on the left is gamma ray (red) and the right track is resistivity (black)

117

APPENDIX C STRATIGRAPHIC CROSS SECTIONS

A series of eight stratigraphic cross sections were made to document the structure of the Elm Coulee field and use that data to build a stratigraphic frame work in Petrel. Refer to chapter 4 for details.

118

A NW

A SE

Figure C1. Stratigraphic cross sections A-A`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A. B NW B SE

Figure C2. Stratigraphic cross sections B-B. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A.

119

C NW

C SE

Figure C3. Stratigraphic cross sections C-C`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A. D SW D NE

Figure C4. Stratigraphic cross sections D-D`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A.

120

E SW

E NE

Figure C5. Stratigraphic cross sections E-E`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A. F SW F NE

Figure C6. Stratigraphic cross sections F-F. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C an d the purple is facies A.

121

H SW

H NE

Figure C7. Stratigraphic cross sections H-H`. The dark gray is the Upper and Lower Bakken Shales, the light gray is facies E and F, the light red is facies B and C and the purple is facies A.

122

APPENDIX D LIST OF WELLS USED IN PETREL MODEL

UWI 25083215160000 25083210250000 25083213650000 25083212060000 25083212780000 25083217460000 25083210880000 25083212720000 25083213530000 25083215010000 25083050070000 25083210110000 25083210180000 25083210210000 25083210220000 25083210230000 25083210270000 25083210310000 25083210360000 25083210380000 25083210520000 25083210570000 25083210790000 25083210810000 25083211250000 25083211330000 25083211460000 25083211540000 25083211670000 25083211690000 25083211800000 25083211810000 25083211850000 25083211870000 25083211980000 25083212030000 25083212090000 25083212130000 25083212180000 25083212240000 WellName TRUE WITT GERALD K LARSEN SOUTHLAND NEVINS COON OBERGFELL CENEX SORENSEN PUTNAM ET AL LADD DUNCAN-JOHNSON PATRICK ALBIN MCMILLEN THE EXPLORATION STATE C J DYNNESON C DYNNESON DYNNESON JOHNSON WOODS DYNNESON "A" DYNNESON PENNZOIL SORENSON WOODS DYNNESON "A" COON MIAMI J.A. COON PUTNAM OBERGFELL KNIGHT & MILLER RASMUSSEN PUTNAM ENSERCH CENEX GARTNER PENNZOIL NEVINS STEINBEISSER CUNDIFF BRINKERHOFF MONTANA PENNZOIL LARSON SIMARD COON BRINKERHOFF BOND PENNZOIL FRANZ LARSON FARMERS UNION STATE ALTON-CHRISTIANSEN SAWYER FARMERS LARSON U. V. OBERGFELL Well # 24-35 1-12 1-19 2-25 3-34 8-17 1-24 1-34 2-34 1-36 1-32 1-32 1-7 4-1-4 1-7 2-7 1-4 2-7 1-1 1-2 1-20 1-21 1-11 2-20 1-23 1-9 31-2 8-22 1-16 1-24 10-27 1-25 1-8 1-8 2-26 5-36 1-9 1-8 15-21 1-34 Sec 35 12 19 25 34 17 24 34 34 36 32 32 7 4 7 7 4 7 1 2 20 21 11 20 23 9 2 22 16 24 27 25 8 8 26 36 9 8 21 34 T 23 23 23 23 23 23 23 23 24 24 24 24 23 23 23 23 23 23 23 23 23 23 23 23 23 23 22 23 23 23 23 23 23 23 23 23 23 23 23 23 R 58 57 57 57 57 58 58 58 57 58 58 58 58 58 58 58 58 58 57 57 57 57 58 57 58 57 58 58 58 57 58 57 58 57 57 58 58 57 58 57 TD 13225 12625 12620 12536 12630 12720 12745 12600 12750 11003 12648 12728 12542 12650 10550 12464 12695 12492 12583 12700 12530 12570 12720 12540 12561 12405 12463 12620 10500 12438 12500 12540 12550 12518 12556 12600 12552 12500 12575 12604 DigitalCurves CALD DRHO GRR LLD LLS NPHI RHOB SPR TEND TENR ; CLLD DT GRR LL MCAL MLL ; CALI DRHO DT GR LLD LLS MSFL NPHI RHOB ; CALD CORR DPHI GRR LLD LLS MSFL NPHI RHOB SPR ; CALD DPHI DRHO GRR LLD LLS MSFL NPHI RHOB SPR ; CAL1R CALD CALS CALSR DPHI_LS DRHO DT GRD GRR GRS LLD LLS ; CALI GR GRR LLD LLS NPHI RES SPR ; CALS DRHO DT GRS LLD LLS NPHI RHOB SFLU SPR TEND TENR TENS ; CALS DRHO DT GRR LLD LLS MSFL NPHI RHOB TEND TENR TENS ; CALD DPHI GRR LLD LLS NPHI TEND TENR ; ASN CALS CLLD DT GR GRN GRR LL LN LT16 MLL NEUT SNP SPR ; CALS CLLD DT GRR LL MLL ; CALD CALN CORR DPHI GRD GRN ILD ILM LL8 MLL RHOB SNP SPR ; CALS CLLD DT GRR LL MLL NPHI ; CALS CILD DT GRS ILD ILM LL8 SNP SPR ; CALD CORR DPHI GRD ILD ILM RHOB SPR ; CALS DT GRR LL MLL RES ; CALD CORR DPHI GRD ILD ILM LL8 RHOB SNP SPR ; CALD CORR DPHI GRD ILD ILM LL8 RHOB SNP SPR ; CALS CLL DT GRS LL SPHI ; CALS DT GRR GURD ; CALI CILD GRN GRR ILD ILM LL8 SNP ; AC CALS CLLD GRS LL SPHI ; AC CLLD GRN GRR LL3 NEUT ; CALS CORR DT GRS LLD LLS MSFL NPHI RHOB TEND TENS ; CALD CORR GRR LLD LLS MSFL NPHI RHOB SPR ; CALD CORR DPHI GRR LLD LLS NPHI RHOB SPR ; CALD CORR DPHI GRN GRR LLD LLS MSFL NEUT NPHI RHOB SPR ; CALS DT GRR LLD LLS MSFL SPR ; CALD CORR DPOR GRR LLD LLS NPOR RHOB SPR ; CALD CORR DPHI GRR LLD LLS MSFL NPHI RHOB ; CALD CORR DPHI GRR LLD LLS MSFL NPHI RHOB SPR ; CALR CORR DPHI GRR LLD LLS MSFL NPHI RHOB ; CALD CORR DPHI GRR LLD LLS MSFL NPHI RHOB SPR ; CALD CORR DPHI GRR LLD LLS NPHI RHOB SPR ; CALD CALR CALS CORR DPHI DT GRD GRR GRS LLD LLS MSFL NPHI ; CALD CALS CORR DPHI DT GRR LLD LLS MSFL NPHI RHOB SPR ; CALD CORR DPHI GRR LLD LLS MSFL NPHI RHOB SPR ; CALD CORR GRR LLD LLS MSFL NPHI RHOB SPR ; CALD CORR GRR LLD LLS NPHI RHOB SPR ; 123

UWI 25083212270000 25083212390000 25083212400000 25083212460000 25083212550000 25083212800000 25083212870000 25083213050000 25083213080000 25083213410000 25083213430000 25083213520000 25083213660000 25083214330000 25083214480000 25083214500000 25083214770000 25083215320000 25083215460000 25083215510000 25083215530000 25083215810000 25083215900000 25083216200000 25083216340000 25083216430000 25083216510000 25083216750000 25083216900000 25083217050000 25083217140000 25083217640000 25083217720000 25083218050000 25083218880000 25083219090000 25083220250000 25083222680000 25083224320000 25083226950000 25083227390000

WellName FARMERS MONTANA STATE U. V. OBERGFELL FARMERS SUNWALL SOUTHLAND PREVOST MULLINS ALBIN APACHE HILL FARMERS CUNDIFF "R" TENNECO CUNDIFF TENNECO OBERGFELL SOUTHLAND ALICE PREVOST OBERGFELL LAURENT C PREVOST FEE PRUET-ORLO C NEVINS RICHARD HIER GULF OBERGFELL "B" LARSON PENNZOIL LARSON SOUTHLAND NEVINS BEREN OBERGFELL TRUE STEINBEISSER DYNNESON LONETREE CREEK DOTSON AMOCO UNIT A PENNZOIL STATE PENNZOIL FRANZ FARMERS SIMBARD SUN NORGAARD UNIT SUN RANCH 4 "J" MISSION SUNWALL SAWYER E. P. STATE E.P. ANDERSON CHRISTENSEN DYNNESON HEADINGTON DYNNESON NANCE LARSON NANCE LARSON PEABODY-MINIFIE LONETREE R R-EDNA FOGHORN-ERVIN

Well # 4-36 2-23 15-15 1-24 1-32 1-33 1-28 8-22 1-28 1-32 1-29 1-33 1-19 1-13 1-30 1-11 1-14 4-19 2-19 1-21 1-1 1-5 12-1 1-34 1-16 2-6 6-27 1-6 1-2 1-8 1-10 12-16 9-26 10-9 12X-5 32X-32 11-26 8-28 14-H 1-13-1 3-HLID

Sec 36 23 15 24 32 33 28 22 28 32 29 33 19 13 30 11 14 19 19 21 1 5 1 34 16 6 27 6 2 8 10 16 26 9 5 32 26 28 26 1 20

T 23 23 23 23 23 24 24 23 23 23 23 23 23 23 23 23 23 23 23 23 22 23 23 24 23 23 23 23 23 23 23 23 23 23 23 24 23 23 24 23 23

R 57 57 58 56 56 57 56 58 57 57 57 57 56 56 56 57 57 58 57 57 58 58 56 58 57 57 58 58 57 58 57 58 58 58 58 58 57 58 56 56 58

TD 12575 12520 12612 12580 12120 12715 12520 12600 12465 12550 12121 12661 12250 12650 12030 12750 12582 12600 12650 12650 12513 12458 12551 12850 12565 12580 10411 12565 12700 12540 12485 10525 12550 10460 12582 12750 12550 12610 14373 12556 10650

DigitalCurves CALD CORR GRR LLD LLS NPHI RHOB SPR ; CALD CORR GRD GRR LLD LLS NPHI RHOB SPR ; CALD CORR GRR LLD LLS MSFL NPHI RHOB SPR ; CALI DRHO DT GR LLD LLS NPHI RHOB RXO ; CALS DPHI DRHO DT GRR LLD LLS MSFL NPHI RHOB ; CALN CORR DPHI GRR LLD LLS MSFL NPHI RHOB SPR ; CALI DRHO DT GR LLD LLS MSFL NPHI RHOB ; CALR DPHI DRHO GRR LLD LLS NPHI RHOB RXO SPR ; CALD CORR DPHI GRR LLD LLS MSFL NPHI RHOB SPR ; CALR CORR DPHI GRR LLD LLS MSFL NPHI RHOB ; CALS DT GRR RD RMLL RS ; CALR GRR NEUT NPHI RD RMLL RS ; AC CALR CN CORR GRS PORD RHOB RILD RILM RMLL ; CALR DPHI DRHO GRR LLD LLS NPHI RHOB RXO ; AC CALS CNCF CORR DEN GRR PORD RD RMLL RS ; CALD DRHO DT GRR LLD LLS MSFL NPHI RHOB TEND TENR TENS ; CALR DPHI DRHO GRR LLD LLS NPHI RHOB RXO ; CALS DRHO DT GRR LLD LLS NPHI RHOB RXO TEND TENR TENS ; CALS DPHI DRHO DT GRR LLD LLS MLL NPHI RHOB ; CALI DRHO DT GR LLD LLS MSFL NPHI RHOB ; CALD DRHO GRR LLD LLS NPHI RHOB RXO TEND TENR ; CALR DPHI DRHO DT GRR LLD LLS NPHI RHOB RXO ; CALD DPHI DRHO GRR LLD LLS MSFL NPHI RHOB ; CALS DPHI DRHO DT GRR LLD LLS NPHI RHOB RXO ; CALD DPHI DRHO GRR LLD LLS MSFL NPHI PEF RHOB ; CALI DRHO DT GR LLD LLS MSFL NPHI PEF RHOB ; CALD DRHO GRR LLD LLS NPHI PEF RHOB RXO TEND TENR ; CALR DRHO DT GRR LLD LLS NPHI PEF RHOB RXO TEND TENR TENS ; CALR DPHI DRHO GRR LLD LLS NPHI RHOB RXO ; CALR DRHO DT GRR LLD LLS NPHI PEF RHOB RXO TEND TENR TENS ; CALX CDL CNS CORR GRR LLD LLS MSFL RHOB ; CALS CNC CORR DEN DT GRR PORD RD RMLL RS ; CALS CORR DT GRR LLD LLS MSFL NPHI PE PEL PES RHOB TEND TENR TENS ; CALR DPHI DRHO GRR LLD LLS NPHI PEF RHOB RXO ; AT10 AT30 AT90 DPHZ DT GRR HCAL HDRA NPOR PEFZ RHOZ ; AF20 AF30 AF60 AF90 DPHZ DT GRS HCAL HDRA NPOR RHOZ ; DPHZ DT1 DT1R DT4P DT4S DTCO DTRP DTRS DTTP DTTS GRD HCAL; DPHZ DT GRR HCAL HDRA NPOR RHOZ RLA4 RLA5 RXOZ ; AIBD AT20 AT30 AT90 DPHZ DT GRR HDRA NPOR PEFZ RHOZ ; AIBD AT20 AT30 AT90 DPHZ DT GR GRR HCAL HDRA NPOR PEFZ ; AT20 AT30 AT90 CALI DPHZ DTCO3 DTSH5 DTSH6 GR GRD GRR HCAL HDRA ;

124