3D FACIES MODELING

BOREHOLE-IMAGE LOG INTERPRETATION AND 3D FACIES MODELING IN THE MESAVERDE GROUP, GREATER NATURAL BUTTES FIELD, UINTA BASIN, UTAH

by Mirna I. Slim

A thesis submitted to the faculty and 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: _____________________ Mirna I. Slim

Approved: ___________________ Dr. Neil F. Hurley Thesis Advisor

Approved: ___________________ Ir. Max Peeters Thesis Advisor Golden, Colorado Date ___________ _______________________ Dr. John Humphrey Acting Department Head, Department of Geology and Geological Engineering

ii

ABSTRACT

Four borehole image logs from wells in the Greater Natural Buttes field, Uinta basin, Utah constitute the data set of this study. The main objective of the research was to study structural and sedimentary features on the image logs to determine: (1) the orientations of the horizontal principal stresses, (2) the fluvial and shoreface facies proportions, and (3) model the subsurface for productivity prediction. Fractures were picked on the image logs and classified into various fracture sets. The orientation of open natural fractures and drilling-induced fractures helps determine the orientation of the present-day maximum horizontal stress (SHmax). Hydraulic fractures occur parallel to this stress orientation. The mean SHmax and SHmin obtained from the 4 wells trend 103 and 17 degrees from north, respectively. Sedimentary features, dip-magnitude, dip-azimuth patterns, and gamma-ray log shapes were studied to determine the environment of deposition of the various sandstone packages. The vertical distribution and proportions of the various facies in well NBU-222 and their dimensions, based on outcrop analogs, provided the input for a 3D facies model. Pseudo-wells were inserted in the model to check the N/G (net-to-gross ratio) and calculate an average number of sand bodies intersected by a well in the modeled area. The average net-to-gross (N/G) percent in the Mesaverde Group and the Castlegate Sandstone was found to be 55% sandstone. For all wells combined, channel, crevasse-splay, and shoreface sandstones formed 33%, 5%, and 10% of the vertical sections, respectively. Coals, lagoonal, and washover sandstones existed in minor amounts. Also, the facies proportions showed that the Upper Mesaverde Group formed in a fluvial environment, the Lower Mesaverde Group contained deposits from fluvial, shoreface, and transition environments, and the Castlegate Sandstone was deposited in a shallow-marine environment. Finally, the facies model showed that, in a depositional system similar to the Mesaverde Group, half of the sand bodies present may be penetrated by more than one well if the wells are drilled at a 1.35-ac (0.005 km2) spacing. A relationship between the numbers of wells in a drilling scenario to the percentage of sand bodies intersected was obtained and helped determine the best scenario to adopt in a

iii

similar fluvial environment. Between 6 and 50% of sand bodies can be intersected when well spacing varies between 40-ac (0.17 km2) and 5-ac (0.02 km2), respectively. We assumed that all the sand facies have the same sand quality. No porosity or permeability data were included in the facies model to quantitatively check the hydraulic connectivity of the sand bodies. Similarly, no production data were used to check the sand volume predicted by the model or the subsurface sand distribution. Therefore, future work should integrate outcrop and core descriptions and production data.

iv

TABLE OF CONTENT ABSTRACT ....................................................................................................................... iii LIST OF FIGURES ............................................................................................................ xi LIST OF TABLES ............................................................................................................ xix ACKNOWLEDGMENT ....................................................................................................xx

CHAPTER 1: INTRODUCTION........................................................................................1 1.1. Research Objectives.....................................................................................................1 1.2. Studied Wells and Data................................................................................................3 1.3. Previous Work .............................................................................................................3 1.4. Methodology ................................................................................................................5 1.5. Research Contributions................................................................................................7

CHAPTER 2: REGIONAL GEOLOGY ............................................................................8 2.1. North America and Global Tectonics ..........................................................................8 2.2. The Rocky Mountain Region Tectonics ....................................................................10 2.2.1. The Ancestral Rocky Mountains ........................................................................10 2.2.2. The Sevier Orogeny ............................................................................................10 2.3.3. The Laramide Orogeny .......................................................................................12 2.2.4. The Wasatch Range ............................................................................................13 2.3. The Cretaceous Geology of Utah...............................................................................13 2.4. The Uinta Basin .........................................................................................................16 2.5. Regional Stratigraphy ................................................................................................18 2.5.1. The Mesaverde Group in the Studied Wells .......................................................24 2.5.2. Mesaverde Group Environment of Deposition ...................................................25 2.6. The Mesaverde Total Petroleum System ...................................................................29 2.6.1. Source Rocks ......................................................................................................29

v

2.6.2. Maturation...........................................................................................................31 2.6.3. Migration.............................................................................................................32 2.6.4. Reservoir Rocks ..................................................................................................32 2.6.5. Traps and Seals ...................................................................................................32 2.7. Summary of the Greater Natural Buttes Field Characteristics...................................32

CHAPTER 3: BOREHOLE IMAGE LOGS....................................................................34 3.1. Formation MicroImager.............................................................................................34 3.2. The Physics of FMI Measurements ...........................................................................44 3.3. FMI Image Processing ...............................................................................................45 3.3.1. Data Load............................................................................................................45 3.3.2. GPIT Survey .......................................................................................................45 3.3.3. ICS Super Caliper Recalibrator ..........................................................................47 3.3.4. BorEID................................................................................................................47 3.3.4.1. EMEX Correction ....................................................................................... 47 3.3.4.2. Data Equalization........................................................................................ 48 3.3.4.3. Depth and Speed Corrections...................................................................... 48 3.3.5. BorNor ................................................................................................................50 3.3.6. BorScale..............................................................................................................50 3.3.6.1. Conductivity Matching ............................................................................... 50 3.3.6.2. Depth Shift .................................................................................................. 53 3.3.7. BorDip.................................................................................................................53 3.3.8. BorView..............................................................................................................57 3.3.9. Data Save ............................................................................................................59 3.3.10. Dip to ASCII .....................................................................................................62 3.4. Log Quality ................................................................................................................62 3.4.1. Gas Entry ............................................................................................................62 3.4.2. Tool Pull..............................................................................................................62 3.5. Fractures and Structural Analysis ..............................................................................65 3.5.1. Open Fractures ....................................................................................................65

vi

3.5.1.1. Open Natural Fractures ............................................................................... 67 3.5.1.2. Drilling-Induced Fractures.......................................................................... 67 3.5.1.3. Borehole Breakouts..................................................................................... 69 3.5.1.4. Borehole Elongations.................................................................................. 72 3.5.2. Healed Fractures .................................................................................................75 3.5.3. Fracture Aperture ................................................................................................75 3.5.4. Fracture Porosity.................................................................................................78 3.5.5. Fracture Total Trace Length ...............................................................................79 3.6. Bedding planes and Sedimentary Analysis................................................................79 3.6.1. Lithology Determination.....................................................................................79 3.6.2. Environments of Deposition ...............................................................................80 3.6.3. GR Signature of Sandstone Facies......................................................................80 3.6.4. Open-Hole Logs..................................................................................................84 3.6.5. Coal .....................................................................................................................84 3.6.6. FMI Signatures of Sandstone Lithologies...........................................................84 3.6.6.1. Dip Pattern .................................................................................................. 87 3.6.6.2. Dip Magnitude ............................................................................................ 87 3.6.6.3. Scour Surfaces ............................................................................................ 87 3.6.6.4. Dip Azimuth Vector Plots........................................................................... 87 3.6.7. Paleocurrent Directions from Image Logs..........................................................89 3.6.8. Structural Dip Removal ......................................................................................90 3.6.9. Cumulative Dip Plot ...........................................................................................92 3.7. Log Presentation ........................................................................................................95 3.7.1. Track One............................................................................................................95 3.7.2. Track Two...........................................................................................................95 3.7.3. Track Three.........................................................................................................97 3.7.4. Track Four...........................................................................................................97 3.7.5. Track Five ...........................................................................................................98 3.7.6. Track Six.............................................................................................................98 3.7.7. Track Seven ......................................................................................................100 3.7.8. Track Eight........................................................................................................102

vii

CHAPTER 4: BOREHOLE IMAGES: FRACTURE INTERPRETATION .............103 4.1. Well One: Natural Buttes Unit 222..........................................................................103 4.1.1. Open Natural Fractures .....................................................................................104 4.1.2. Drilling-Induced Fractures................................................................................106 4.1.3. Borehole Breakouts...........................................................................................106 4.1.4. Healed (Resistive) Fractures .............................................................................109 4.2. Well Two: Bonanza 4-6 ...........................................................................................109 4.2.1. Open Natural and Partially Healed Fractures ...................................................109 4.2.2. Drilling-Induced Fractures................................................................................111 4.2.3. Borehole Breakouts...........................................................................................113 4.2.4. Healed (Resistive) Fractures .............................................................................113 4.3. Well Three: Pawwinnee 3-181.................................................................................113 4.3.1. Open Natural and Partially Healed Fractures ...................................................115 4.3.2. Drilling-Induced Fractures................................................................................118 4.3.3. Borehole Breakouts...........................................................................................118 4.3.4. Healed (Resistive) Fractures .............................................................................123 4.4. Well Four: Kennedy Wash Federal Unit 16-1 .........................................................123 4.4.1. Open Natural and Partially Healed Fractures ...................................................123 4.4.2. Drilling-Induced Fractures................................................................................125 4.4.3. Borehole Breakouts...........................................................................................125 4.5. Summary ..................................................................................................................127 4.6. Discussion ................................................................................................................127

CHAPTER 5: BOREHOLE IMAGES - FACIES INTERPRETATION ....................134 5.1. Well One: Natural Buttes Unit 222..........................................................................135 5.1.1. Structural Dip Removal ....................................................................................135 5.1.2. Cumulative Dip Plot .........................................................................................138 5.1.3. Dip Azimuth Vector Plot ..................................................................................140 5.1.4. Facies Examples from Well NBU-222 .............................................................143 5.1.5. Facies Proportions.............................................................................................154

viii

5.2. Well Two: Bonanza 4-6 ...........................................................................................154 5.2.1. Structural Dip Removal ....................................................................................154 5.2.2. Cumulative Dip Plot .........................................................................................157 5.2.3. Dip Azimuth Vector Plot ..................................................................................160 5.2.4. Facies Examples from Well Bonanza 4-6.........................................................160 5.2.5. Facies Proportions.............................................................................................166 5.3. Well Three: Pawwinnee 3-181.................................................................................174 5.3.1. Structural Dip Removal ....................................................................................174 5.3.2. Cumulative Dip Plot .........................................................................................177 5.3.3. Dip Azimuth Vector Plot ..................................................................................177 5.3.4. Facies Examples from Well Pawwinnee 3-181 ................................................182 5.3.5. Facies Proportions.............................................................................................193 5.4. Well Four: Kennedy Wash Federal Unit 16-1 .........................................................193 5.4.1. Structural Dip Removal ....................................................................................198 5.4.2. Cumulative Dip Plots........................................................................................198 5.4.3. Dip Azimuth Vector Plot ..................................................................................202 5.4.4. Facies Examples from Well KWFU 16-1.........................................................202 5.4.5. Facies Proportions.............................................................................................209 5.5. Discussion ................................................................................................................209

CHAPTER 6: 3D FACIES MODEL ...............................................................................218 6.1. Model Grid Design ..................................................................................................220 6.2. NBU-222 Facies Proportions...................................................................................226 6.3. Petrel Input: Sand Element Dimensions and Shapes ...............................................226 6.4. Petrel Model Results ................................................................................................232 6.5. Discussion ................................................................................................................240

CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS.................................243 7.1. Conclusions..............................................................................................................243

ix

7.2. Recommendations....................................................................................................244

REFERENCES..................................................................................................................246 APPENDIX A: SUMMARY OF FACIES INTERPRETATION.................................CD APPENDIX B: MAPVIEWS OF WELL-SPACING SCENARIOS AND CONNECTIVITY ANALYSIS................................................................CD APPENDIX C: WELL DATA .........................................................................................CD

x

LIST OF FIGURES Figure 1 - 1: The Greater Natural Buttes field in the Uinta basin........................................2 Figure 1 - 2: Map view showing the four studied wells ......................................................4 Figure 2 - 1: Physiographic provinces of the Uinta-Piceance basin region.......................11 Figure 2 - 2: The main structural elements that resulted from the Laramide Orogeny in Utah.....................................................................................................14 Figure 2 - 3: Late Cretaceous Paleogeography of the western United States....................15 Figure 2 - 4: Outcrops of Upper Cretaceous rocks and the main uplifts in Utah and surrounding states............................................................................................17 Figure 2 - 5: Generalized stratigraphic column of Uinta-Piceance province.....................19 Figure 2 - 6: Divisions and nomenclatures of the different sandstones of the Mesaverde Group...................................................................................................20 Figure 2 - 7: Fluvial environment from borehole image logs............................................26 Figure 2 - 8: A model of nearshore sand deposits .............................................................26 Figure 2 - 9: Depositional model showing the sedimentary structures in (A) longitudinal bar and (B) transverse bar in a braided channel ................................27 Figure 2 - 10: Depositional model showing the sedimentary structures in (A) a mid- or downstream point bar and (B) upstream end of a point bar......................28 Figure 2 - 11: Sedimentary features of a prograding shoreface succession with (A) high energy and (B) low energy ......................................................................30 Figure 3 - 1: Different generations of borehole image tools..............................................35 Figure 3 - 2: Schlumberger‘s FMI tool ..............................................................................36 Figure 3 - 3: A pad and a flap mounted on each of the caliper arms .................................37 Figure 3 - 4: The coverage of various borehole image tools .............................................39 Figure 3 - 5: Microresistivity curves recorded by each pad...............................................43

xi

Figure 3 - 6: The sequential image processing chain in Geoframe....................................46 Figure 3 - 7: Depth correction to align the microresistivity curves ...................................49 Figure 3 - 8: Static and dynamic normalization.................................................................51 Figure 3 - 9: Static vs. dynamic image ..............................................................................52 Figure 3 - 10: A schematic diagram showing the dip and azimuth of a bedding plane.......................................................................................................................54 Figure 3 - 11: Cross-correlations between the FMI pads...................................................56 Figure 3 - 12: Dipping bedding plane traced with a sine wave .........................................58 Figure 3 - 13: Structural cross section built using the monocline model...........................60 Figure 3 - 14: Structural cross section built using the similar-fold model.........................61 Figure 3 - 15: Gas entry in well Bonanza 4-6....................................................................63 Figure 3 - 16: Tool pull in well Pawwinnee 3-181 ............................................................64 Figure 3 - 17: Fracture morphologic types ........................................................................66 Figure 3 - 18: Dynamic image from well Bonanza 4-6 showing an open fracture and cross stratifications ............................................................................68 Figure 3 - 19: Drilling-induced fractures ...........................................................................70 Figure 3 - 20: Borehole breakouts and drilling-induced fractures from well Bonanza 4-6 ...........................................................................................................71 Figure 3 - 21: Asymmetric elongation in well section (A) with and (B) without borehole breakouts .................................................................................................73 Figure 3 - 22: Fracture classifications and their orientation with respect to the maximum and minimum stress directions .............................................................74 Figure 3 - 23: Healed fractures with the halo effects.........................................................76 Figure 3 - 24: FMI response to conductive fractures.........................................................77 Figure 3 - 25: GR curve signatures in different fluvial facies ...........................................81 Figure 3 - 26: Tadpole patterns and GR signatures for continental shelf ..........................82

xii

Figure 3 - 27: Tadpole analysis and GR signatures in continental and shelfdelta........................................................................................................................83 Figure 3 - 28: The GR curve signature of the braided-channel facies of Figure 2-9 ..........................................................................................................................85 Figure 3 - 29: The GR curve signature of the meandering system bars ............................86 Figure 3 - 30: Different dip patterns ..................................................................................88 Figure 3 - 31: A dip-azimuth vector plot of (A) all shale beds in well NBU222 and (B) sections of opposed accretions...........................................................88 Figure 3 - 32: A Schmidt plot prepared using Geoframe for shale beds in well NBU-222................................................................................................................91 Figure 3 - 33: Cumulative dip plot of the shale beds in well NBU-222 ............................93 Figure 3 - 34: Slumps and sedimentary deformation in well Pawwinnee 3-181 ...............94 Figure 3 - 35: The first five tracks of the FMI PDS log presentation................................96 Figure 3 - 36: The sixth track of the FMI log presentation................................................99 Figure 3 - 37: Tracks seven and eight of the FMI log presentation.................................101 Figure 4 - 1: The open natural fracture set in well NBU-222..........................................105 Figure 4 - 2: The drilling-induced fracture set in well NBU-222 ....................................107 Figure 4 - 3: The borehole breakout set in well NBU-222 ..............................................108 Figure 4 - 4: The open natural and partially healed fracture sets in well Bonanza 4-6 .........................................................................................................110 Figure 4 - 5: The drilling-induced fracture and borehole breakout sets in well Bonanza 4-6 .........................................................................................................112 Figure 4 - 6: The healed fracture set in well Bonanza 4-6...............................................114 Figure 4 - 7: The open natural and partially healed fracture sets in the shallower interval of well Pawwinnee 3-181.......................................................116 Figure 4 - 8: The open natural fracture set in the deeper interval in well Pawwinnee 3-181.................................................................................................117

xiii

Figure 4 - 9: The drilling-induced fracture and borehole breakout sets in the shallower interval in well Pawwinnee 3-181 .......................................................119 Figure 4 - 10: The drilling-induced fracture and borehole breakout sets in the deeper interval in well Pawwinnee 3-181............................................................120 Figure 4 - 11: The healed fracture set in the shallower interval of well Pawwinnee 3-181.................................................................................................121 Figure 4 - 12: The healed fracture set in the deeper interval of well Pawwinnee 3-181 ....................................................................................................................122 Figure 4 - 13: The open natural fracture set in well KWFU 16-1....................................124 Figure 4 - 14: The drilling-induced fracture and borehole breakout sets in well KWFU 16-1 .........................................................................................................126 Figure 4 - 15: (A) Vertical extension fractures in the MWX well, and strike directions of (B) open natural fractures and (C) drilling-induced fractures from the 23 wells in the Piceance basin................................................130 Figure 4 - 16: Section of the geologic map of the Vernal area in Utah ...........................131 Figure 5 - 1: Stereonet plot of the “Sedimentary_Dip” set in well NBU-222 .................136 Figure 5 - 2: Stereonet plot of the “Bed_Boundary” dip set in well NBU-222 ...............137 Figure 5 - 3: Cumulative dip plot of the bedding planes in well NBU-222.....................139 Figure 5 - 4: Cross section showing the shale bedding planes in well NBU-222............141 Figure 5 - 5: Dip azimuth vector plot of all shale beds in well NBU-222.......................142 Figure 5 - 6: Channel-sand fill in well NBU-222 ............................................................144 Figure 5 - 7: DAVP of the sand package seen in Figure 5-6 ...........................................145 Figure 5 - 8: Point bar from well NBU-222.....................................................................147 Figure 5 - 9: Section from well NBU-222 of a sand package from the Lower Mesaverde ............................................................................................................148 Figure 5 - 10: DAVP of the sand layers seen in Figure 5-9.............................................149 Figure 5 - 11: Channel-fill deposits from well NBU-222................................................150

xiv

Figure 5 - 12: The DAVP of amalgamated sand of Figure 5-11 .....................................151 Figure 5 - 13: Two sand packages from well NBU-222..................................................152 Figure 5 - 14: The DAVP of both sand layers of Figure 5-13. ........................................153 Figure 5 - 15: Stereonet plot of the various bedding plane dip sets in well BON 4-6...............................................................................................................156 Figure 5 - 16: Cumulative dip plot of the bedding planes in well BON 4-6....................158 Figure 5 - 17: Cross section showing the shale bedding planes in well BON 46............................................................................................................................159 Figure 5 - 18: DAVP of all shale beds in well BON 4-6 .................................................161 Figure 5 - 19: Fining-upward meandering channel fill in well BON 4-6 ........................162 Figure 5 - 20: The DAVP of the sand in Figure 5-19 ......................................................163 Figure 5 - 21: Washover sand in well BON 4-6 ..............................................................164 Figure 5 - 22: The DAVP of the lagoon washover sand of Figure 5-21..........................165 Figure 5 - 23: Washover fan/lagoonal shoreline sandstone from well BON 4-6.............167 Figure 5 - 24: DAVP of the washover fan seen in Figure 5-23 .......................................168 Figure 5 - 25: Washover fan sand in well BON 4-6 ........................................................169 Figure 5 - 26: The DAVP of the washover fan sand seen in Figure 5-25........................170 Figure 5 - 27: A thick amalgamated shoreface sand section in the Lower Sego Formation in well BON 4-6 .................................................................................171 Figure 5 - 28: Vector plot of a lower shoreface sand in well BON 4-6...........................172 Figure 5 - 29: Stereonet plot of the various bedding plane sets of the shallower interval of well Pawwinnee 3-181 .......................................................................175 Figure 5 - 30: Stereonet plot of the various bedding plane sets in the deeper interval of well Pawwinnee 3-181 .......................................................................176 Figure 5 - 31: Cumulative dip plot of the bedding planes in well Pawwinnee 3181........................................................................................................................178

xv

Figure 5 - 32: Cross section showing the shale bedding planes of the upper interval of well Pawwinnee 3-181 .......................................................................179 Figure 5 - 33: Cross section showing the shale bedding planes of the lower interval of well Pawwinnee 3-181 .......................................................................180 Figure 5 - 34: The DAVP of the bedding planes picked in shale sections in the (A) shallower interval and (B) deeper interval of well Pawwinnee 3181........................................................................................................................181 Figure 5 - 35: Fluvial deposits in the Upper Mesaverde in well Pawwinnee 3181........................................................................................................................183 Figure 5 - 36: DAVP of the fluvial sand in Figure 5-35..................................................184 Figure 5 - 37: Amalgamated fluvial channel deposits in well Pawwinnee 3181........................................................................................................................185 Figure 5 - 38: DAVP of the sand section seen in Figure 5-37.........................................186 Figure 5 - 39: Washover fan from well Pawwinnee 3-181..............................................187 Figure 5 - 40: DAVP of the washover fan sand of Figure 5-39.......................................188 Figure 5 - 41: Shoreface sand seen in the Lower Sego Formation in well Pawwinnee 3-181.................................................................................................189 Figure 5 - 42: DAVP of the shoreface sand seen in Figure 5-41.....................................190 Figure 5 - 43: Upper shoreface sand from the Castlegate Formation in well Pawwinnee 3-181.................................................................................................191 Figure 5 - 44: DAVP shows episodes of progradation and retrogradation......................192 Figure 5 - 45: Planar cross stratification in a silty sand package from the Mancos Shale in well Pawwinnee 3-181 .............................................................194 Figure 5 - 46: DAVP of the silty sand in the Mancos Shale seen in Figure 5-45............195 Figure 5 - 47: Turbidite deposits from the Mancos Shale in well Pawwinnee 3181........................................................................................................................196 Figure 5 - 48: Stereonet plot of the various bedding plane sets in well KWFU 16-1 ......................................................................................................................199 Figure 5 - 49: Cumulative dip plot of the bedding planes in well KWFU 16-1 ..............200

xvi

Figure 5 - 50: Cross section showing the shale bedding planes in well KWFU 16-1 ......................................................................................................................201 Figure 5 - 51: DAVP showing an inflection point at 10,231ft.........................................203 Figure 5 - 52: DAVP of the shale bedding planes in well KWFU 16-1 ..........................204 Figure 5 - 53: Fining-upward fluvial sand package from the Upper Mesaverde Group in well KWFU 16-1 ..................................................................................205 Figure 5 - 54: DAVP of the fluvial sand in Figure 5-53..................................................206 Figure 5 - 55: Fining-upward fluvial sand from the Upper Mesaverde Group in well KWFU 16-1..................................................................................................207 Figure 5 - 56: DAVP of the fluvial sand in Figure 5-55..................................................208 Figure 5 - 57: Braided stream sand deposits from the Upper Mesaverde Group in well KWFU 16-1 .............................................................................................210 Figure 5 - 58: DAVP of the braided stream channel fill of Figure 5-57..........................211 Figure 5 - 59: Shoreface sand section from the Castlegate Formation in well KWFU 16-1 .........................................................................................................212 Figure 5 - 60: DAVP of the shoreface sand in Figure 5-59 .............................................213 Figure 5 - 61: Facies proportions in all 4 wells ...............................................................216 Figure 5 - 62: Facies proportions calculated from all 4 wells in (A) MVU, (B) MVL, and (C) the CGATE Formation.................................................................217 Figure 6 - 1: A map view to show the locations of the four wells in Petrel ....................219 Figure 6 - 2: Well NBU-222 facies log............................................................................221 Figure 6 - 3: Surfaces created to connect the formation tops ..........................................222 Figure 6 - 4: Polygon that designates the area modeled around well NBU-222..............223 Figure 6 - 5: (A) The horizons relating formation tops and separating (B) the modeled zones......................................................................................................224 Figure 6 - 6: Example of the input data in Petrel.............................................................227 Figure 6 - 7: Elliptical shape used to model sand bodies in Petrel ..................................230

xvii

Figure 6 - 8: Petrel 3D facies model ................................................................................233 Figure 6 - 9: A map view showing the 34 wells distributed at 5-ac (0.02 km2) spacing around well NBU-222.............................................................................234 Figure 6 - 10: Number of sand bodies intersected in different infill-drilling scenarios...............................................................................................................238 Figure 6 - 11: Relationships of the number of wells in a drilling scenario to (A) the number and (B) percentage of the sand bodies intersected .....................239 Figure 6 - 12: Facies proportions in MVU and MVL of well NBU-222.........................242

xviii

LIST OF TABLES Table 1 - 1: Four studied wells and their locations in the Uinta basin.................................5 Table 2 - 1: Greater Natural Buttes Field summary...........................................................33 Table 3 - 1: Generations of borehole image tools..............................................................38 Table 3 - 2: FMI Specifications .........................................................................................41 Table 3 - 3: FMI Applications ...........................................................................................42 Table 3 - 4: Cumulative dip calculations spreadsheet .......................................................92 Table 3 - 5: Dip sets, their color codes, and tadpole symbols used on the FMI log presentation....................................................................................................100 Table 4 - 1: Structural analysis summary ........................................................................128 Table 5 - 1:Formation tops in the studied wells...............................................................135 Table 5 - 2: Facies proportions in well NBU-222............................................................155 Table 5 - 3: Facies proportions in well BON 4-6.............................................................173 Table 5 - 4: Facies proportions in well Pawwinnee 3-181...............................................197 Table 5 - 5: Facies proportions in well KWFU 16-1 .......................................................214 Table 6 - 1: The coordinates, KB elevations, and TD of the four studied wells..............225 Table 6 - 2: Sand element orientations and dimensions used for the modeled zones ....................................................................................................................228 Table 6 - 3: Elements/facies proportions in 5-ac (0.02 km2) spaced wells around well NBU-222..........................................................................................235 Table 6 - 4: Number of the sand bodies penetrated by 34 wells distributed in a 5-ac (0.02 km2) grid spacing around well NBU-222 ...........................................236 Table 6 - 5: Summary of the well-spacing scenarios, the wells they contain, and the number of sand bodies intersected in each..............................................239

xix

ACKNOWLEDGMENT

Many people have contributed to make my research and 2 years at the Colorado School of Mines an interesting experience. I acknowledge the Kerr McGee Corporation and the Center of Petrophysics for funding my research and study. I would like to thank my advisor, Dr. Neil F. Hurley, for his guidance, support, revision, and comments on my work. I feel lucky to have known a person like you that can teach a lot, even outside a classroom. It was a privilege to work with you and be your student. Thanks go to my co-advisor, Professor Max Peeters, for his support and help during my last semester. Also, special thanks go to Dr. Donna Anderson for her time and insights to better understand fluvial sandstones and their architecture. Thanks go to Dr. Piret Plink-Bjorklund for her comments and revisions. I extend my special thanks to Jerry Cuzella at Kerr McGee Corporation, who helped provide the logs and the required data to build the Petrel model. I feel grateful to Mrs. Charlie Rourke for many things. Charlie, your smile and kindness made it possible to go through tough days. Even during the days when I did not need to bother you to help me with something, it was nice to know that you are around and always ready for assistance. Thanks go to Dr. David Pyles for the time we spent discussing sedimentary structures in fluvial systems and Ms. Shauna Gilbert for her patience and effort to keep the computer laboratories in top shape. I would like to thank all of the geoscientists at Schlumberger’s Data and Consulting Services at Greenwood, Colorado. Looking back at all the work I did in your office, I think that this research would have taken much more time without the great support from Jim Urdea during my internship with Schlumberger in 2005 and long after. Jim, thanks for the permission to get access to Geoframe and Petrel. Also, this research would have been much more difficult without Randy Koepsell. Randy, you are a great person and mentor. Thanks for your time and the commitment you showed to teach me Geoframe. You made my work on image logs and Geoframe a pleasant and learning

xx

experience. Thanks go to Adil Manzoor, Tom Becker, Steve Sturm, and Jim Benson. I would like to thank Barbara Luneau and Andie Gomez for their time and help with Petrel. My Petrel modeling would have been nearly impossible without the help of Ekeng Henshaw. Also, I would like to thank Leonardo (Leo) Vega for writing a quick code that helped me count facies and sand bodies in my wells. Thanks go to Dr. Matthew Pranter for his guidance and time spent to discuss the Petrel facies model. I would like to thank Nicholas Sommer for his time discussing the connectivity application and the workflow document he prepared and shared with me. Thanks go to Dr. Zulfiquar Reza who provided the connectivity application. Thank you for your time, comments, feedback, and revisions of my work. Thanks go to Jennifer Brotherhood and Janice Roberts, from ExxonMobil Exploration Company, for their help during my internship in 2006. Finally, I want to express my appreciation to my family and friends. Your love, understanding, and support kept me going away from all of you. I would like to thank Maya for the many and long “therapy sessions” on the phone. Also, I would like to thank all the friends I made in Colorado. Knowing all of you made me feel at home. You made me enjoy this great state and, specifically, the outdoors and tearing/repairing an ACL.

xxi

For my uncle and friend Akram M. Arawi.

xxii

CHAPTER 1

INTRODUCTION

Many of the Rocky Mountain hydrocarbon fields are known for their low permeability, on the order of micro and millidarcies. The Greater Natural Buttes (GNB) field in the Uinta basin (Figure 1-1) is no exception. The low permeability in many fields makes production, even that of gas, a challenging task. The low permeabilities are aggravated because the Upper Cretaceous Mesaverde Group has fluvial and shoreface deposits that are often non-continuous and lenticular in shape. Thus, it is important to better understand the occurrence of the sand bodies, their dimensions, and orientations in order to model the subsurface for production prediction and efficiency. Wireline image logs, although less commonly acquired compared to other conventional open-hole logs, prove to be a good structural and stratigraphic tool to interpret and evaluate the Upper Cretaceous sandstone reservoirs in the Mesaverde Group. Image logs are useful to interpret sand bodies in terms of various elements/facies in a certain environment of deposition. Also, referring to studies done on sandstone outcrops of the Mesaverde Group in the Piceance basin, the dimensions and orientation of various sand facies can be used to model an area around one of the study wells.

1.1. Research Objectives The research has two main objectives. First, the structural elements picked in the wells (mainly fractures and borehole breakouts) are studied to determine the orientations of the principle stress directions in the GNB field. The results are compared to the stress map of North America (NA) and other work carried out to determine the principle stress directions in other fields in the basin (Chapter 4). Second, the stratigraphic aspect of the study aims to:

1

Figure 1 - 1: The Greater Natural Buttes field in the Uinta basin. USA map after www.enchantedlearning.com; Uinta basin map modified after Johnson and Roberts (2003).

2

1- Classify the sand bodies into various fluvial and shoreface facies or elements, and find their proportions in every well and in different zones in each well. 2- Assign facies/element dimensions based on outcrop analogs. 3- Build a subsurface 3D facies model to reflect the occurrence, distribution, orientations, and dimensions of the sand bodies. 4- Show how various infill-drilling scenarios can intersect different numbers of sand bodies, and draw conclusions on the best drilling scenario(s) to adopt for optimum production efficiency assuming that all sand facies have the same quality.

1.2. Studied Wells and Data The Uinta basin (Figure 1-1) is an asymmetric foreland basin in NE Utah. The basin is an extension of the Piceance basin in NW Colorado. It covers an area of 9,300 mi2 (24,087 km2) and is a “typical Rocky Mountain asymmetrical basin” (Osmond, 1965; Osmond, 1968). The study area is the GNB field in the eastern part of the Uinta basin (Figure 1-1). Four different fields (Table 1-1): the Bonanza (BON), the Kennedy Wash Federal Unit (KWFU), the Natural Buttes (NBU), and the Pawwinnee (PAW) fields are represented in the study. They are located in the central part of the Uinta basin, inside the GNB field. The studied data are four Formation Micro-Imager (FMI) logs from wells in the above mentioned fields (Table 1-1; Figure 1-2).

1.3. Previous Work There have been numerous publications to study the principal stress orientation and stratigraphy in North America (NA) and, specifically, the Uinta-Piceance region. Lorenz (2003) presented a relationship between fractures in the Mesaverde Group in the Piceance basin and regional geological events. Zoback and Zoback (1989) classified the Rocky Mountain (RM) region into various stress “provinces” and published a stress map of the continental United States. Zoback et al. (1985) studied the relation between in-situ

3

Figure 1 - 2: Map view showing the relative locations of the four studied wells. No field boundaries are specified on the map.

4

Table 1 - 1: Four studied wells and their locations in the Uinta basin.

stress and borehole breakouts, and Koepsell et al. (2003) interpreted FMI images of 23 different wells in the Piceance basin and showed the strike direction distribution of open natural fractures. Luthi (2001) and Grace and Newberry (1998) presented detailed work on how to understand, process, and interpret image logs and any feature seen on them. Longman and Koepsell (2005) published detailed work done on many wells (including the ones studied in this research) in the GNB field where image logs were used to interpret environments of deposition of various sandstone formations. The Upper Cretaceous stratigraphy was intensively studied by Johnson (2003), Hettinger and Kirschbaum (2003), and Lawton (1986) who described outcrops for sedimentary structures and paleocurrent direction determinations. The architecture and dimensions of the sand bodies in various sandstone formations of the Mesaverde Group were published by Cole and Cumella (2003; 2005) and Anderson (2005) and studied by Panjaitan (2006) and German (2006). Petrel facies modeling done by Henshaw (2005) was reviewed as an example for fluvial facies modeling in the Uinta basin.

1.4. Methodology The following is the methodology used to meet the study objectives, with details presented in subsequent chapters:

5

1- Process the FMI logs using Schlumberger’s Geoframe software to make them ready for interpretation (Chapter 3). 2- Trace key structural elements (faults, micro-faults, open natural fractures, resistive/healed fractures, borehole breakouts, and drilling-induced fractures) and assign dip and dip azimuth for each. Classify each type of feature into a set, and determine average dip and dip azimuth for each set to infer principal stress directions (Chapter 4). 3- Compute cumulative fracture trace lengths and fracture porosities in each well (Chapter 4). 4- Trace key sedimentary features (bedding planes, cross bedding, erosional scours, and slumps) on the image logs, classify them into dip sets, and assign each sedimentary feature a dip and dip azimuth (Chapter 5). 5- Interpret sedimentary features and open-hole logs to determine and differentiate sand elements into shoreface sand, channel sand/point bars, or crevasse splays (Chapter 5). 6- Filter shale bedding planes from the rest of the data and sort them into dip domains to prepare cumulative dip plots. The change in the plot slope gradient may help to predict faults, micro-faults, and sequence boundaries that may not be clear from the image logs (Chapter 5). 7- Determine, from different shale intervals, the post-depositional dip/structure and remove it. 8- Construct dip azimuth vector plots to determine, when possible, accretion orientations in the channel sands and point bars and infer paleocurrent directions (Chapter 5). 9- Determine the proportion of every facies element in each well (Chapter 5). 10- Create a synthetic facies log in Petrel and use facies dimensions and orientations to populate a conceptual 3D model in the vicinity of well NBU-222 (Chapter 6). Well NBU-222 was chosen because its FMI image log has less gas entry and tool pull than the remaining logs. Also, we could locate, around well NBU-222, other wells with formation tops specified. This provided more control to shape the horizons separating the modeled zones in Petrel. 11- Study various infill drilling scenarios to determine the number of sand bodies intersected in each scenario, and predict the effects of each on production and its

6

efficiency (Chapter 6).

1.5. Research Contributions The main results of the research include the determination of the directions of the principle stress orientation in 4 wells. The present-day maximum horizontal stress direction (SHmax) was found to trend 103 degrees from north. The minimum horizontal stress (SHmin) was found to trend 17 degrees from north. The orientations agree with the regional trend. Stratigraphic analysis shows that the logged intervals of the Mesaverde Group contain sand and shale in almost equal proportions. The average net-to-gross (N/G) percent in the studied wells in the Mesaverde (Upper and Lower) and the Castlegate Sandstone was found to be 55%. The sand was divided into crevasse splays, channel sands, and shoreface sands that formed 5%, 33%, and 10% of the logged intervals in the mentioned formations, respectively. Lagoonal and washover sands were encountered as well. We could infer from the results that the Upper Mesaverde Group was deposited in a fluvial environment, the Lower Mesaverde Group contained deposits from fluvial, shoreface and transition environments, and the Castlegate Sandstones was deposited in a shallow-marine environment. The facies model built in Petrel showed that, in a depositional system similar to the one we are studying, all the sand bodies present can be penetrated when drilling wells at a 1.35-ac (0.005 km2) spacing; half of sand bodies may be penetrated by more than one well. When we vary the well spacing between 20-ac (0.08 km2) and 5-ac (0.02 km2), the percentage of sand bodies intersected will vary between 12% and 50%, respectively, of the total sand bodies found in the subsurface. To drill at 40-ac (0.17 km2) spacing or larger may bypass more than 90% of the subsurface sand bodies.

7

CHAPTER 2

REGIONAL GEOLOGY

This chapter is a review of the global tectonics of North America (NA), the regional tectonics of Utah and the Uinta basin, and the regional stratigraphy of the UintaPiceance region. A summary of the global tectonics that formed and shaped the craton of NA will be summarized with a focus on the main regional tectonic events that shaped the Rocky Mountain (RM) region and formed the structural features surrounding the Uinta basin. The regional stratigraphy of interest will be summarized with a focus on the producing sandstones of the Upper Cretaceous Mesaverde Group and their environments of deposition. Finally, the Mesaverde Group will be presented as a petroleum system in the Greater Natural Buttes (GNB) field area, reviewing the petroleum system elements and summarizing the GNB characteristics.

2.1. North America and Global Tectonics The North American craton, known as Laurentia, has been coherent since 1.7 Ga and included Greenland and parts of Scotland until their separation during the Late Cretaceous. It is a group of “seven micro-continents,” called provinces, collected by orogenic activity between 2.0 and 1.8 Ga (Hoffman, 1989). Some of the provinces belonged to subducting plates during convergence or forelands during collisional orogenies. Others were parts of overriding plates at one point of their history and hinterlands during orogenies. Early Proterozoic rifting and collisional deformation controlled the dimensions of these provinces (Hoffman, 1989). The northwestern margin of North America witnessed Precambrian orogenies (Grenville and Racklan orogens) that were the result of “thrusting directed toward the interior of Laurentia.” The Proterozoic ended with a continental breakup taking place at 0.8 Ga, which developed a number of aulacogens, one of which is the Uinta Mountains

8

(Hoffman, 1989). Bally et al. (1989) reviewed the main tectonic events that shaped NA and influenced its evolution to the configuration that we know today. A western volcanic arc was present through the Paleozoic and Mesozoic. During the mid-Paleozoic, the western margin of NA was involved in major episodes of shortening that resulted in continent-directed thrust belts with a “locally well-developed foredeep succession containing Devono-Carboniferous sediments” (Oldow et al., 1989). During the Late Permian (255 Ma), the western half of the supercontinent Pangea (Euramerica, Gondwana, and Siberia) formed after the ocean basin separating the first from the latter two closed. The final reconstruction of this supercontinent took shape by late Carboniferous and remained unaltered until the Early Jurassic (Bally et al., 1989). The western margin type of NA changed from being a passive margin to an active one during the Permian and Early Triassic (Oldow et al., 1989; Johnson, 1992). During the Middle and Upper Triassic and the Early Jurassic, a new rift system formed along the east coast of NA. This rift system was the precursor of a subsequent system that led to the development of the Gulf of Mexico and the Central Atlantic Ocean. The first appearance of a narrow oceanic basin was seen at the beginning of the Cretaceous (144 Ma), and later the “Proto-Caribbean Ocean formed,” separating NA from South America (SA) (Late Cretaceous). At about 119 Ma (Late Neocomian to early Aptian), the Central Atlantic Ocean showed slight expansion and a small oceanic basin appeared, separating South America from Africa. During the Late Cretaceous (84 Ma, late Santonian), sea-floor spreading continued between Africa and NA. A new rifting system started developing between NA and Europe. The Triassic to Paleocene evolution of the Cordillera created strike–slip faults that displaced previously overthrust terranes, accretionary wedges, and foreland fold belts, 3,106 mi (5,000 km) in length (Oldow et al., 1989). This created an estimated 62-124 mi (100-200 km) of shortening in the RM region (Oldow et al., 1989). The early Oligocene (38 Ma) witnessed the opening of the Labrador Sea, the formation of Baffin Bay, and the separation between Greenland and Norway. Also, the leading edge of the Pacific Plate started to impinge upon the western margin of NA, causing the San Andreas fault to develop and the Gulf of California to open. This change in plate motion and stress distribution caused many tectonic changes in the western

9

United States. The craton of the Southwestern part of NA was involved in local basement uplifts of either the Paleozoic Wichita-Ancestral Rocky Mountain system or the Laramide Southern Rocky Mountains (Bally et al., 1989).

2.2. The Rocky Mountain Region Tectonics The main tectonic events in the Rocky Mountain Region will be summarized below.

2.2.1. The Ancestral Rocky Mountains The Ancestral Rocky Mountains (ARM) formed during a period ranging from Pennsylvanian to Early Permian between orogenies related to the Ouachita-Marathon fold belt. They represent the youngest Paleozoic structural event (Hintze, 1988; Oldow et al., 1989), and many of their structural deformations have been overprinted by the Laramide deformation. The main uplifts, believed to have been driven by the collision of NA and SA, are the Pennsylvanian Uncompahgre and Front Range uplifts of New Mexico and Colorado (Johnson, 1992). Both were reactivated during the Laramide deformation in the southern RM. Also, subsidence led to the formation of the Eagle, the northern Paradox, and the SE Oquirrh basins, the southern Wyoming shelf, and the eastern Callville shelf (Figure 2-1). Triassic strata covering the Uncompahgre Highlands marked the end of the ancestral Rockies uplifts. Deposits were accumulating (thickening to the west) from the Triassic through the Early Cretaceous (Hintze, 1988).

2.2.2. The Sevier Orogeny The Sevier Orogeny took place between 100-80 Ma (Stokes, 1986) and was characterized by eastward thin-skinned (piggy-back) thrusting in response to active subduction along the west margin of NA (Hintze, 1988; Johnson, 2003). The orogenic

10

Figure 2 - 1: Physiographic provinces of the Uinta-Piceance basin region. The Wasatch line is the eastern edge of the Wasatch Plateau and Range; GNB = Greater Natural Buttes. Modified after Raisz (1972) and Stokes (1977) in Johnson (1992).

11

belt crosses Utah diagonally. With the ARM completely eroded, the sea flooded Utah from the east causing a westward-trending shoreline to become parallel to the “Meso-Cordilleran” foothills. The Cordillera became higher and wider, and the region east of the Wasatch line started to sink. Both uplift and subsidence, due to thrust loading and perhaps mantle flow as “magmatism began to migrate inboard into the central Rocky Mountain region” as a result of the decrease in angle of subduction of the Farallon Plate during the Campanian (Decelles, 2004), were extreme. This produced a great amount of sediments to be deposited as well as rapid transport by streams and created the foreland basin east of the thrust belt (Johnson, 2003). The orogeny reached its “maximum intensity during Late but not latest Cretaceous, had its maximum effects west of the Wasatch line or along it,” and ceased to move by the end of the Cretaceous or very shortly thereafter (Stokes, 1986).

2.3.3. The Laramide Orogeny The Laramide Orogeny is a thick-skinned thrust event with thrust faults extending deeply into the basement rocks (Johnson, 2003). This “intraforeland” basement uplift disrupted the pattern of subsidence in the foreland region and shifted the sediment accumulation eastward (Decelles, 2004). It has affected territory east of the Wasatch line (Stokes, 1986), and overlapped with the Sevier compression (Hintze, 1988) and continued right after it, from the mid-Late Cretaceous (Campanian) (80-40 My) to the early Eocene (84-66 Ma) (Stokes, 1986). The uplift was caused by a few degrees of rotation of the Colorado Plateau relative to the continental interior when the eastern Pacific floor was being subducted beneath NA (Hintze, 1988). The basement uplift was the result of a shear zone that extended into the lower crust and by medium-angle (25-35 deg) reverse faults, with no mantle involvement (Bally et al., 1989). The direct results of this orogeny are the Rocky Mountains and the uplift of the Uinta Mountains, which, geomorphologically, are an east-west trending anticlinal fold belt (160 mi [257.5 km] long and 30 mi [48 km] wide). Other minor uplifts resulted as well, such as the San Rafael Swell, Circle Cliffs, Monument, Kaibab, and Uncompahgre uplifts (Figure 2-2). In addition to uplifts, the Laramide Orogeny created mini fluvio-

12

lacustrine basins by disrupting the large foreland basin created during the Late Cretaceous through the Eocene (Johnson, 2003; Stokes, 1986). The reason why the Laramide deformation is concentrated in the central RM is perhaps its occurrence during a time of faster than usual convergence of the Farallon and the NA plates (Johnson, 2003) as well as the “the Late Cretaceous-Early Tertiary flatslab” event (Decelles, 2004). Subsidence during the Early Cretaceous was induced by thrust loading by adjacent thrust sheets (Johnson, 2003; Bally et al., 1989). The Uinta basin was created by a “deep trough formed to the south of the southward-thrusting Laramide Uinta uplift” (Johnson, 2003).

2.2.4. The Wasatch Range The Wasatch Range does not belong to any of the major orogenies mentioned above. It is a Tertiary mountain range believed to have originated after and caused by the Sevier Orogeny (Stokes, 1986). Other disturbances affected Utah. These were regional uplifts known towards the end of the Cretaceous-early Tertiary period (Stokes, 1986). In conclusion, Tertiary uplifts affected all of NA, causing the Cretaceous sea to withdraw on a regional scale (this is supported by the exposure of marine strata in the Grand Canyon of Arizona) (Stokes, 1986).

2.3. The Cretaceous Geology of Utah Four major physiographic provinces occur in the State of Utah: the Colorado Plateau, the Middle Rocky Mountains, the Basin and Range, and the Colorado Plateau/Basin-Range (Stokes, 1986). The state is crossed by the NE-SW Wasatch line (Figure 2-1) that was a late Precambrian rift arm that widened to become the “Paleozoic Cordilleran Geosyncline” (Stokes, 1986). The Cretaceous lasted for 78 My and was the time “of the last epicontinental sea in Utah” (Hintze, 1988; Osmond, 1965) (Figure 2-3). During the later phase of the

13

Figure 2 - 2: The main structural elements that resulted from the Laramide Orogeny in Utah. An approximate location of the GNB (Greater Natural Buttes) field is indicated. Modified after Stokes (1986).

14

Figure 2 - 3: Late Cretaceous (~75 Ma) Paleogeography of the western United States. UT = Utah and CO = Colorado. Modified after Blakey (2003).

15

Cretaceous, a major marine-flooding event, from north and south, divided NA into two large land masses. Towards the end of the period, the oceans withdrew from the continental interior as a result of both local and global influences. The local ones are indicated by the belt of deltas and nearshore deposits that migrated progressively eastward. The global influence was the fact that continental seas withdrew from major continents at almost the same time (Stokes, 1986). Most of Utah’s Cretaceous rocks were deposited during the second half of the period as the prevalent flat topography (Latest Jurassic and early Cretaceous) led to mild erosion. The sediments were being deposited in both central and eastern Utah, creating a wedge of sediment reaching up to 3,000 ft (914 m) in thickness to be later eroded by the Colorado River during the Cenozoic. Cretaceous deposition (Figure 2-4) in the eastern part of Utah was accompanied by subsidence to create a total Cretaceous sediment thickness exceeding that of the remaining Paleozoic and Mesozoic sediments (Hintze, 1988).

2.4. The Uinta Basin The Uinta basin (Figure 1-1) was part of a Cretaceous foreland basin which is “a continent-long area of downwarping that stretched from the Arctic to Mexico” (Johnson and Roberts, 2003). The basin, 100 mi (161 km) long and nearly 100 mi (161 km) wide, is asymmetrical and bordered by the Uinta Mountains to the north, the San Rafael Swell to the southwest, and the Wasatch Mountains to the west (Spencer 1987; Osmond, 1968). It extends into the Paradox and Piceance basins to the south and east, respectively, with the Douglas Creek Arch bounding it to the east (Osmond, 1968). The basin has occupied an “intraplate geologic setting throughout Phanerozoic time and is mostly underlain by Phanerozoic strata” (Johnson, 1992). Thermally driven subsidence began at 590 Ma. The development of the Antler orogenic belt to the west resulted in an increase in subsidence rate and a change in depositional pattern in the westernmost Uinta-Piceance basin (Johnson, 1992). The whole period was one of submergence and subsidence in the more stable central and eastern parts of the region.

16

Figure 2 - 4: Outcrops of Upper Cretaceous rocks (green patches) and the main uplifts in Utah and surrounding states. Modified after Johnson (2003).

17

From Mid-Late Mississippian to early Early Permian, the overlapping influences of tectonic interactions along continental margins to the west and southeast led to a complex mix of contractional, extensional, and strike-slip deformation in the region. Erosion of the Antler belt started, and a complex heterogeneous tectonic style, characterized by local uplift, subsidence, and volcanism, was the result. The evolution of the basin during the period extending from Middle Jurassic to Early Cretaceous was a continuation of an “initial pulse of rapid, asymmetric subsidence linked to a thrusting event in eastern Nevada followed by a period of slower, more uniform subsidence and relative tectonic quiescence” (Johnson, 1992). During the Late Cretaceous, the region was “located near the western shoreline of the Western Interior seaway and within the Cretaceous Rocky Mountain Foreland basin” (Hettinger and Kirschbaum, 2003). The seaway retreated from the Uinta-Piceance region during the latest Cretaceous (Turonian to Early Campanian), fluctuated during the Campanian with an orientation between N65E and N15W, and was outside of the area completely by the Maastrichtian (Hettinger and Kirschbaum, 2003). The foreland basin was segmented by basement-cored uplifts into restricted lacustrine basins (Johnson, 1992). With the Laramide Orogeny, compressional forces controlled the structural development of the region of the Uinta basin caused by the Uinta (north) and the Uncompahgre Uplifts (south) (Cuzella and Stancel, 2006). The basin, as we know it today, took shape as a result of relatively higher uplift of its margins than its interior, creating 3,000-6,000 ft (914-1,829 m) of relief. This development started during the Early Tertiary (Paleocene or Eocene) and was intermittent since then (Osmond, 1965). The early Tertiary, therefore, is the time for the formation of the Uinta basin, with its northern boundary formed during the Precambrian with the abrupt rise of the Uinta Mountains (Osmond, 1965; Hintze, 1988).

2.5. Regional Stratigraphy The stratigraphic interval of interest extends from the marine Mancos Shale to the Green River Formation, excluding both (Figure 2-5). The latter contains the main producing sandstones in the Uinta basin and the ones penetrated by the studied wells (Cuzella and Stancel, 2006) (Figures 2-5 and 2-6).

18

Figure 2 - 5: Generalized stratigraphic column of Uinta-Piceance province. Modified after Johnson (2003).

19

Figure 2 - 6: Different divisions and nomenclatures of the different sandstones of the Mesaverde Group. The red rectangle indicates the GNB field. The interval bounded by the blue lines is referred to as the CGATE Sandstone in the text. The Lower Mesaverde interval (MVL) is indicated by the green segment. Modified after Hettinger and Kirschbaum (2003).

20

The Upper Cretaceous Mesaverde Group has a thickness ranging from 2,000 to 3,000 ft (607-914 m), and it can reach 4,000 ft (1,219 m) of “interbedded sandstones, siltstones, shale and coal,” as demonstrated by Longman and Koepsell (2005) through well data. The group was sourced from eroded sediment from the Sevier orogenic belt to the west. The sediments were shed to central and eastern Utah during repetitive cycles of regressions and transgressions of the Campanian sea (Hintze, 1988; Johnson, 2003) in the “Western Interior Seaway” (Cole and Cumella, 2003). The sediment wedge thickened (thinning toward the east and north) as subsidence took place. The sandstones of the Mesaverde Group vary in thickness from 2 to 50 ft (0.6-15 m) and can be found at depths ranging from 6,000 to 13,000 ft (1,829-3,962 m) (Longman and Koepsell, 2005). The group was divided differently by various authors in different basins and within different parts of the same basin. The various groupings of the different sandstones of the Mesaverde Group are shown in Figure 2-6. In an ascending order, the formations are: 1- The Mancos Shale is black marine shale. On the logs, gas entries are commonly seen coming from the shale sections. The Mancos “B” interval is “a laterally extensive package of thin, relatively clean, very fine grained sandstones in the Mancos Shale.” It forms the deepest interval penetrated in the wells (Longman and Koepsell, 2005). 2- The Blackhawk Formation is a sheet-like sandstone deposited in wave-dominated deltas and along strandlines downcurrent from the deltas. The change and progradation in the shoreline, distributary channel switching, and local and regional transgressions produced stacked and imbricated sequences of delta-front sandstones. The organic-rich deposits that are normally present behind the Blackhawk Formation shoreline thin remarkably in the Natural Buttes field (Pitman et al., 1987). 3- The Castlegate Sandstone is a “massive sandstone body ranging between 150-260 ft (46-79 m) and deposited in a braided plain environment” (Cuzella and Stancel, 2006). In general, this sandstone does not constitute main producing reservoirs in the GNB (Cuzella and Stancel, 2006). The top of the Blackhawk Formation is not identified in the wells, and the Castlegate Sandstone discussed to in the text includes upper sections of the Blackhawk Formation (Figure 2-6).

21

4- The Buck Tongue is a shale member of the Mancos Shale deposited 68 Ma. It is “the last preserved marine incursion” in the GNB field (Cuzella and Stancel, 2006; Longman and Koepsell, 2005). 5- The Sego Sandstone is a “shallowing-upward marine shoreline sandstone” (Longman and Koepsell, 2005). It was deposited in tidal channel, estuarine and shoreface settings (Cuzella and Stancel, 2006). 130 ft (40 m) thick, the sandstone is rarely completed in the GNB field as it lacks a lateral seal (Cuzella and Stancel, 2006). In the studied wells, the Sego Sandstone is a transitional interval containing mainly washover sand and shale, deposited in a lagoonal environment (Longman and Koepsell, 2005). The blocky Sego Sandstone is a shaly lower-shoreface sandstone that grades up into upper shoreface and marine sandstone. This sand has good reservoir characteristics but is mostly wet. 6- The Neslen Formation is a continental coal-bearing formation which is restricted to the eastern part of the Uinta basin (Pitman et al., 1987). It was deposited in coastal plains and floodplain swamps (Longman and Koepsell, 2005). It has a gradational and intertonguing contact with the underlying marginal-marine Sego Sandstone (Pitman et al., 1987). The Corcoran Member, the Cozette Member, and the Rollins Member of the Iles Formation are equivalent to the Lower Mesaverde and the Neslen Formation in the Uinta basin (Johnson, 1993; Cumella and Ostby, 2003; Koepsell et al., 2003; Cole and Cumella, 2005). These sand members were deposited in inner-shelf, deltaic, shoreface, estuarine, and lower coastal plains settings (Cole and Cumella, 2005). The trend of the Corcoran and Cozette shorelines was NE-SW, whereas the Rollins shoreline trended NNE-SSW (Cumella and Ostby, 2003; Koepsell et al., 2003). The change in the trend may be the result of a “tectonically influenced shift in the basin subsidence” (Cumella and Ostby, 2003; Johnson, 1993). Various regressive cycles have led to the intertonguing of the Lower Mesaverde Sandstones with the Mancos Shale (Johnson, 2003). This indicates that marine flooding happened very rapidly at the end of each regressive cycle, with the Castlegate Sandstone being the oldest regressive cycle in the Mesaverde. The shoreline is believed to have moved from the NW to SE during those regressions, resulting in a purely fluvial environment that dominated the area after the regressions (Johnson, 2003). In the studied wells, the

22

Neslen Formation is a “relatively homogeneous, non-coaly package of sandstones and shales…with consistent very low, non-spiky resistivity response. It contains 4-10 fluvial channel sandstones less than 15 ft (4.6 m) thick.” The Lower Neslen interval contains coal beds and rare thin sandstones deposited in a coastal plain (Longman and Koepsell, 2005). 7- The Farrer and Tuscher Formations, referred to as the Upper Mesaverde Group, are not differentiated in the GNB field (Cuzella and Stancel, 2006). In the Piceance basin, both sandstones are grouped into the Williams Fork Formation that is conformable with the overlying Iles Formation. The Williams Fork Formation was deposited in a non-marine environment, on a coastal plain “west of a prograding shorelines” (Cumella and Ostby, 2003; Johnson, 1993). In the Piceance basin, the formation has an upper sand-rich interval and a lower sand-poor interval (Cole and Cumella, 2005), and a variable thickness due to truncation and variations in subsidence during deposition. It consists of interbedded sandstone, mudrock, and coal deposited in alluvial-plain, coastal-plain, and marginal-marine settings. In its lower interval, it contains many coal seams in the “Cameo-Wheeler coal zone (Figure 2-6), which consists of interstratified coal, carbonaceous to very carbonaceous mudrocks, ironstone concretion, sandstone, and rare conglomerate,” as described from cores and outcrops in the Piceance basin (Cole and Cumella, 2005). The Cameo coal zone was deposited in “paludal environments of the lower coastal plain” (Cumella and Ostby, 2003).

Above the Mesaverde, younger formations were penetrated by wells with no image log coverage. These are the Wasatch Formation and the Green River Formation. The former is 2,000-3,000 ft (610-914 m) thick, deposited in a marginal lacustrine environment. It is comprised of shale, siltstones, paludal limestones, and lenticular fluvial sandstones (Cuzella and Stancel, 2006). The Green River Formation is mainly shale interbedded with sandstone and limestone, deposited in a lacustrine setting (Cuzella and Stancel, 2006). The lacustrine shale of the Green River Formation (the base of which is likely time transgressive) forms a seal against upward migration of the gas from the

23

Mesaverde Group. This same shale is a good marker to separate the Mesaverde Group from the Wasatch Formation.

2.5.1. The Mesaverde Group in the Studied Wells Longman and Koepsell (2005) presented a concise summary of the formations and environments of deposition encountered in the wells they studied (including our four wells) in the GNB field. They divided the wells into zones without necessarily defining accurately the formation tops. These intervals are the Castlegate Formation, the Sego Formation, the Neslen Formation, the Mesaverde Group (divided into Lower, Middle, and Upper Mesaverde), and the Mancos Shale. The guidelines used to divide the Mesaverde Group, according to Longman and Koepsell (2005), are the presence of the thick sandstone packages with blocky GR vs. the thin fining-upward fluvial channel sandstones. Also, the presence and absence of coal beds were another criterion followed to divide the Mesaverde Group as follows (Longman and Koepsell, 2005):

1- The Upper Mesaverde has braided stream channels and a blocky GR signature. 2- The upper interval of the mid-Mesaverde contains fining-upward fluvial channel fills (very few with blocky GR) with thick floodplains. No coal is present. 3- The Middle Mesaverde is recognized by its thick braided-stream sandstones with blocky GR signature. The sandstones can be separated by thin intervening shale beds. This is the most sand-rich part of the Mesaverde. 4- The lower fluvial interval of the Mesaverde contains fining-upward fluvial channel sandstones.

For this study, however, the formation tops in the wells, used to create horizons for modeling in Petrel, were adopted as provided by the Kerr McGee Corporation. Two main zones, the Upper Mesaverde and the Lower Mesaverde, are modeled in well NBU222 (Chapter 6). The lower zone includes what Longman and Koepsell (2005) interpreted to be the upper Castlegate Formation, the Sego Sandstone, and the Neslen Formation.

24

The upper zones include what Longman and Koepsell (2005) interpreted to be “fluvial” and “braided” deposits of the undifferentiated Tuscher and Farrer Formations.

2.5.2. Mesaverde Group Environment of Deposition A sedimentary environment is a “geographically restricted part of the earth’s surface, which can be distinguished….by the set of physical, chemical, and biological processes which exist and characterize it.” A facies “is the sum of physical, chemical, and biological characteristics which permits to differentiate a sedimentary body from another” (Serra, 1985). The main environments of deposition encountered in the studied wells are the fluvial and shoreface environments. Both environments existed during the Upper Cretaceous when the Mesaverde Group was being deposited in the GNB field. Sandstones from both a fluvial environment (Figure 2-7) and a shoreface environment (Figure 2-8) were seen in the studied wells. The fluvial environment is a continental depositional environment where sediments are carried and transported by running water to be later deposited into lacustrine or marine basins (Galloway and Hobday, 1996). The stream systems develop on different slopes with segments having low sinuosity (braided streams) or high sinuosity (meandering streams). Fluvial facies, according to Galloway and Hobday (1996), can be a combination of channel fill (lag, accretionary channel, different types of bars), channel-margin deposits (crevasse splays, levees), and floodbasin deposits (floodplain, backswamp, and interchannel lakes). The internal structure in each facies depends on the geometry of the channel and the direction of water flow. Downstream bed accretion causes deposition of longitudinal bars (Figure 2-9), which are characteristic of a braided-stream system. Lateral accretion and the formation of point bars are the main characteristics of high-sinuosity, meandering-stream systems (Figure 2-10). Tabular and trough cross stratification may be rare to common in the middle to upper point-bar succession. Also, a vertical decrease in grain size is observed. The top of an old point bar is usually vegetated and capped by fine floodplain sediments (Galloway and Hobday, 1996). Scour surfaces are common in a fluvial system. Channel lags, coarse bed-loads, and plant debris may be seen on top of those basal erosional surfaces.

25

Figure 2 - 7: Fluvial environment showing different facies identified from borehole image logs. Modified after Selley (1978) in Hallam (1981).

Figure 2 - 8: A model of nearshore sand deposits showing the locations where shoreface sands form. Modified after Scholle and Spearing (1982) in Prothero (1990).

26

Figure 2 - 9: Depositional model showing the sedimentary structures in (A) longitudinal bar and (B) transverse bar in a braided channel. Modified after Galloway and Hobday (1996).

27

Figure 2 - 10: Depositional model showing the sedimentary structures in (A) an upwardfining sequence in a mid- or downstream point bar and (B) upstream end of a point bar. Modified after Galloway and Hobday (1996).

28

Meandering and braided streams were both dominant, and different sections of the Mesaverde Group have channel fills of each stream type (Chapter 5). Several authors (Longman and Koepsell, 2005; Patterson et al., 2003) have interpreted many sandstone packages to be braided-stream deposits in the Mesaverde Group of both the Uinta and Piceance basins. A shoreface environment is a marine environment that develops on the seaward side of a barrier or in the open coast (Friedman and Sanders, 1978). The shoreface extends from low-tide level to wave-base level (30-40 ft; 9-12 m) (Figure 2-8). In other words, the shoreface extends from the outer edge of the beach to a distinct change in slope. Sediments are mainly deposited by wave action (Friedman and Sanders, 1978). The lower-shoreface deposits accumulate at the break in slope where the shoreface grades into the shelf. Bioturbation, if seen on the image logs, is a good indication of a lowershoreface environment. The sediment size and dominant physical and biogenic structures depend on the sediments available as well as the wave-energy regime. In high-waveenergy regimes, parallel-laminated sand is deposited “under conditions of intense bottom shear and may preserve low-amplitude undulations or hummocky cross-stratifications.” In the upper shoreface (inner surf zone), sediments are dominated by “onshore, offshore, and longshore” currents; longshore-directed trough cross beds are characteristic. Also, onshore-dipping planar cross-beds are deposited by bar migration (Figure 2-11).

2.6. The Mesaverde Total Petroleum System The following is a summarized review of the elements of the Mesaverde total petroleum system (TPS).

2.6.1. Source Rocks In the Uinta basin, the Mesaverde Group is sourced mainly “by coal and associated organic-rich strata” (Johnson and Roberts, 2003) like the “carbonaceous and coaly shales interbedded with the sandstones (Pitman et al., 1987). The coals and 29

Figure 2 - 11: Sedimentary features of a prograding shoreface succession with (A) high energy and (B) low energy. After Galloway and Hobday (1996).

30

carbonaceous shales are believed to have accumulated in mires, swamps, and marshes associated with deltaic and coastal-plain environments. Gas is sourced from the coal zones in the sandstones, especially in the lowermost formations (the Neslen Formations) and has migrated either vertically or up-dip into sandstone reservoirs in the upper units of the group (Johnson and Roberts, 2003). Pitman et al. (1987) reported total organic carbon (TOC) values from the Neslen and the Blackhawk Formations in the GNB field. The values range from 1.30 to 5.04% and 1.00 to 8.59%, respectively. The source zones in the Neslen Formation have “relatively low hydrogen and oxygen indices (128-254 mg HC/g TOC, 5-26 mg CO2/g TOC)” (Pitman et al., 1987). The Blackhawk Formation, although deeper and more mature than the Neslen Formation, shows similar hydrogen and oxygen indices (Pitman et al., 1987).

2.6.2. Maturation Knowing that the onset of thermogenic gas generation is between 0.73-0.75%, the coal-bearing zones of the Mesaverde Group have exceeded the level of thermal maturity throughout much of the Uinta basin (Johnson and Roberts, 2003). “The onset of this thermal maturity in the basin began 42 Ma with peak gas generation occurring between 26 and 17 Ma” (Johnson and Roberts, 2003). Cooling hindered gas generation that is believed to be occurring today only in zones where the Mesaverde Group temperature exceeds 200 ºF” (Johnson and Roberts, 2003). The hydrocarbon-generation history was evaluated on samples from the Blackhawk and Neslen Formations in the eastern Uinta basin (Pitman et al., 1987). The authors reported vitrinite reflectance values that range from 0.72% to 0.82% (mean standard deviation of 0.03) in the Neslen Formation in the Natural Buttes field and from 0.89% to 1.01% (mean standard deviation of 0.06) in the Blackhawk Formation in the same field (Pitman et al., 1987). The present day geothermal gradient in the area of Natural Buttes is estimated to be about 29ºC/km (Pitman et al., 1987).

31

2.6.3. Migration The capacity of coal to store gas is high, although it decreases with increasing temperature and decreasing pressure (Johnson and Roberts, 2003). The accumulations in the Uinta basin could have begun to form during the gas-generation peak (26 Ma). The expelled gas from the coal zones is believed to have migrated vertically through fractures to higher sand (Cretaceous and Tertiary) formations as lower intervals became saturated (Johnson and Roberts, 2003).

2.6.4. Reservoir Rocks The producing (and potential) gas reservoirs are mainly the fluvial sandstones of the Mesaverde Group and the (Lower Tertiary) Wasatch Formation. These are mainly lenticular sand bodies. The reservoirs are tight with permeability ranging between 0.001 and 0.1 md (Johnson and Roberts, 2003). The average porosity in the Mesaverde Group is 11% (Cuzella and Stancel, 2006).

2.6.5. Traps and Seals The trapping mechanism in the Uinta basin is largely stratigraphic and diagenetic; the discontinuous nature of the sandstone formations and their low permeability contribute most to the trapping. Compartmentalization is enhanced by lateral facies changes, minor fault offsets, and variations in diagenesis (Johnson and Roberts, 2003). The lacustrine shale layer of the Green River Formation acts as a regional seal (Johnson and Roberts, 2003).

2.7. Summary of the Greater Natural Buttes Field Characteristics A summary of the GNB field characteristics is presented in Table 2-1 from Cuzella and Stancel (2006). The table lists many characteristics of the field, the

32

Mesaverde Group reservoirs, and the produced gas, as well as various exploration and production practices applied to the field.

Table 2 - 1: Greater Natural Buttes Field summary. After Cuzella and Stancel (2006).

33

CHAPTER 3

BOREHOLE IMAGE LOGS

Borehole image logs are electronic images of the borehole wall obtained by measuring the resistivity variation of the rocks and fluids as well as the contrast in the resistivity measurement between different layers or structural features crossing the borehole. The resistivity contrast on the image is represented by different colors, with dark colors representing low resistivity measurements and bright colors representing high resistivity measurements. The formation resistivity is a function of the shale content, lithology, and the fluid content in pores and fractures (Grace and Newberry, 1998). The technique of acquiring borehole images is based on dipmeter technology available commercially since the 1950s (Hurley, 2004). The main objective of the dipmeter measurement is to obtain structural dip and azimuth (defined later) of any plane or planar feature traversed by the borehole (Luthi, 2001). The technology has evolved from having tools with three arms and three buttons to sophisticated imaging and dipmeter measurements (Table 3-1 and Figure 3-1).

3.1. Formation MicroImager The borehole image logs used in this study were acquired using Schlumberger’s FMI (Formation MicroImager) tool. The FMI is an open hole logging tool that provides microresistivity measurements and is logged in wells drilled with water-based muds. It is one of the various generations of borehole imaging tools (Figure 3-1). The FMI (Figure 3-2) consists of an electronic cartridge, which contains all the tool electronics and acts as an electrode for current return, and an insulated mechanical sub or sonde that terminates with two 2-arm calipers at 90º to one another. Each of the arms has a pad and a “flap” mounted on it (Figure 3-3). Both the flap and pad are free to tilt independently of the toolstring itself (Rider, 1996). The curvature of the pad and the

34

Figure 3 - 1: Different generations of borehole image tools. The SHDT is Schlumberger’s Stratigraphic High Resolution Dipmeter Tool. Number in parentheses represents the year in which the tool was released. Modified after Hurley (2004) and Grace and Newberry (1998).

35

Figure 3 - 2: Schlumberger‘s FMI tool. The toolstring shows the cartridge, the insulated sub, and a closer view of the calipers. After Schlumberger (2002).

36

Figure 3 - 3: A pad and a flap mounted on each of the caliper arms. Note the offset between the two rows of electrodes on each pad and flap, and the offset between the pad and flap. After Schlumberger(2002) and Rider (1996).

37

Table 3 - 1: Different generations of borehole image tools with different names and service companies.

hinged design of the flap allow both of them to better adhere to the borehole wall. Both the flap and the pad will be referred to as pads for the rest of the chapter. Each pad contains 24 microresistivity electrodes arranged in two rows 0.3 in (0.75 cm) apart (Figure 3-3) (Rider, 1996), and, thus, the tool has a total of 192 electrodes or buttons. Each of the electrodes has an effective diameter of 0.2 in (5 mm). This dimension is the vertical resolution of the tool, and any feature smaller than 0.2 in (5 mm) cannot be seen by the FMI (Grace and Newberry, 1998). The arrangement of the electrodes and the offset of the two rows (Figure 3-3) insure that the azimuthal sampling rate is the same as the vertical sampling rate (Luthi, 2001). The FMI has a very shallow depth of investigation. Its signal penetration is 0.55 in (1.4 cm) (Rider, 1996). The offset between the flap and pad on each arm, as well as the 0.1 in (0.25 cm) offset of the rows (Figure 33), allows the FMI image to cover 80% of the borehole wall in an 8 in (20.3 cm) hole (Figure 3-4). The FMI can be run in boreholes ranging in diameter from 6.25 in (15.88 cm) to 21 in (53.34 cm), with a maximum logging speed of 1800 ft/hour (549 m/hour). However, the preferred logging speed during image acquisition is 1600 ft/hour (488 m/hour). This gives a high sampling rate with the tool acquiring one sample every 0.1 in (2.5 mm) of vertical displacement. The calipers, magnetometers, and accelerometers are sampled every 1.5 in (3.8 cm) (Rider, 1996). The FMI has a “very large dynamic range,” and it can measure resistivities ranging from less than 0.1 ohm-m to more than 10,000 ohm-m. Also, it has a low

38

Figure 3 - 4: A chart showing the coverage of various borehole image tools in wells of different diameters. Modified after Grace and Newberry (1998).

39

sensitivity to heavy mud, borehole ovalization, and rugosity (Grace and Newberry, 1998). The tool maximum pressure and temperature ratings are 20,000 psi and 350º F (175º C), respectively (Schlumberger, 2002). Table 3-2 is a summary of the FMI specifications and Table 3-3 is a summary of the FMI applications. The FMI is a combinable tool and can be connected to other open-hole tools. However, it is always the bottom-most tool being terminated by a “bottom nose” on the caliper arms. The optimum conditions in which an FMI log can be acquired is to follow the logging speed limits, have the mud resistivity (Rm) not exceeding 50 ohm-m (waterbased mud), and have the ratio of the formation resistivity to Rm less than 1,000. Otherwise, the sharpness of the image will be affected as the current will flow in the borehole instead of the formation. Also, centralization of the tool is needed in deviated boreholes (deviation exceeding 10º). This is done by applying pad pressure on the caliper arms (Grace and Newberry, 1998). A General Purpose Inclinometry Tool (GPIT) is needed with every FMI run. A GPIT contains a 3-axis accelerometer and 3 magnetometers. The former gives acceleration data which helps in the speed correction and in determining the exact position of the tool. The magnetometer determines the tool orientation with respect to true north by measuring the declination, which is the angle between true north and magnetic north (Grace and Newberry, 1998). To compute the dip of any planar feature seen on an image, the following is needed: 1) The relative position of 3 points on the plane. 2) Orientation of the tool knowing the azimuth of pad #1 (P1AZ) as the magnetometer measures the angle between pad #1 and the magnetic north. 3) Angle and direction of deviation of the borehole/tool.

During an FMI logging run, the tool is always assumed to be moving at the same speed of the cable supporting it, except when run in sticky holes. In this case, the speed of the tool will accelerate or decelerate, and speed correction is needed to eliminate the error in the depth offset in the microresistivity curves (Figure 3-5). Speed correction is done using the accelerometer data.

40

Table 3 - 2: FMI Specifications. After Schlumberger (2002).

41

Table 3 - 3: FMI Applications. After Schlumberger ( 2002). 1- Structural Geology • Structural dip • Faults 2- Sedimentary Features • Sedimentary dip • Paleocurrent direction • Sedimentary bodies and their boundaries • Anisotropy, permeability barrier and paths • Thin-bedded reservoirs 3- Rock Texture • Qualitative vertical grain-size profile • Carbonate texture • Secondary porosity • Fracture systems 4- Complement to whole core, sidewall core and formation tester programs • Depth matching and orientation for whole cores • Reservoir description of intervals not cored • Depth matching for sidewall core samples and MDT (Modular formation Dynamic Tester) probe settings 5- Geomechanical Analysis • Drilling induced fractures • Calibration for Mechanical Earth Modeling • Mud weight selection 6- Geology and Geophysics workflow • Deterministic reservoir modeling • Distribution guidance for stochastic modeling • Realistic petrophysical parameters

42

Figure 3 - 5: A schematic diagram showing a four-pad image tool with microresistivity curves recorded by each pad. Note the shift in a bed signature on each of the curves. The shift helps determine the dip magnitude of the bed. After Grace and Newberry (1998).

43

The FMI can be comparable to an oriented core and has a wide range of applications (Table 3-2). The images are used for sedimentologic and structural interpretation. They provide input for geomechanical and reservoir modeling.

3.2. The Physics of FMI Measurements The electrical resistivity image of the FMI is acquired by having a low-frequency EMEX (Emetteur d’Excitation) signal used to focus a “rapidly changing high frequency signal” emitted from the pads (Rider, 1996). This technique is known as “passive focusing” (Luthi, 2001). The pad faces are equipotential surfaces (as the mud around them is at the same potential) relative to the return electrode, which is the tool electronic cartridge itself (Rider, 1996; Luthi, 2001). The insulating material covering the sonde separates the 2 electrode sections (pads and cartridge) and prevents the current injected by the electrodes from flowing through the tool body. This tool design forces the current emitted from the buttons to penetrate the formation at a right angle. The current injected from each of the pad electrodes is a function of the formation resistivity in front of it (Luthi, 2001), and the changes in the formation resistivities (facing the pads and buttons) cause changes in the current density (Rider, 1996). The button arrays sample the current density across the pads. The signal recorded by an FMI consists of two components: a deep signal and a shallow signal. The former is “provided by the total current” injected and is equivalent, in depth, to a resistivity measurement acquired by a laterolog tool (Luthi, 2001). The latter is “provided by the variations of the button current” and is equivalent to a microlog (Luthi, 2001). Each of the buttons records a microresistivity curve, and then an average microresistivity curve is computed for the whole pad (Figure 3-5). The total of 192 curves recorded by the buttons form the “matrix which is processed into the FMI image” (Rider, 1996).

44

3.3. FMI Image Processing Data processing is needed in order to convert the raw acquisition data (microresistivity curves measured by each electrode) into a visual representation or an image (Grace and Newberry, 1998; Rider, 1996). Discussed below is the processing done on the four FMI logs that constitute the data set of our study. Schlumberger’s Geoframe software was used for data processing and image interpretation. Geoframe has a set of geological modules built in it; each with a set of functions. Several of those modules were used, and a chain was built linking them together (Figure 3-6) in a manner to have the data output of one module (higher in the chain) acting as data input of another module (lower in the chain). The sequential steps processing the data and using the various Geoframe modules will be discussed in this section.

3.3.1. Data Load “Data Load” is the first module used to load the raw data into Geoframe. Multiple log runs can be loaded separately if the final log presentation is to contain curves and data from each. The data loaded for each well were DLIS (digital log interchange standard) files obtained from logging jobs consisting of two separate runs: the Platform Express (PEX) run as well as FMI. PEX is a Schlumberger open-hole tool consisting of GR (gamma ray), porosity, density, and resistivity tools in a combined toolstring.

3.3.2. GPIT Survey The “GPIT Survey” module is used to check whether the GPIT’s accelerometer and magnetometers were properly working during the FMI log run. This module cannot correct or recompute faulty GPIT data acquired during the logging job. It is used to check the validity of the GPIT-recorded data. Without good GPIT data, the dipmeter measurements will be useless and the FMI image and any feature on it will not be oriented.

45

Figure 3 - 6: The sequential image processing steps built as a processing chain in Geoframe. Each square contains the name of the Geoframe module used. The hierarchy of the modules uses the output of one (higher in the chain) module to become the input of the other (lower in the chain). (Geoframe Software Package; Courtesy of Schlumberger).

46

3.3.3. ICS Super Caliper Recalibrator The FMI has two 2-arm calipers at a 90º angle to one another. Two orthogonal borehole diameter values, referred to as caliper readings, are acquired when the caliper arms are fully opened. The caliper curves (noted as C1 and C2) are the distances between arms 1 and 3, and arms 2 and 4. In the “ICS Super Caliper Recalibrator” module, the caliper reading can be checked to see if any offset and/or gain needs to be applied. Usually, during a logging run, the FMI is logged shortly in the well casing to record caliper readings in the casing itself. Then, the readings are compared to the known casing inner diameter. Offsets are applied, if needed, to have the calipers read the casing inner diameter. Once the offset is applied, both C1 and C2 curves are recomputed. Note that a different offset may be needed for each caliper. The importance of having accurate caliper readings is the fact that they are used when borehole deviation is computed. Also, the caliper readings reveal the borehole geometry, which is an important parameter used in the geometrical correction to account for pad positions when doing the speed correction (Grace and Newberry, 1998).

3.3.4. BorEID BorEID is the first module that starts processing the FMI raw data. The following is an overview of all the functions done when running this module.

3.3.4.1. EMEX Correction: The first step in processing is to correct for the EMEX variation. This is needed because during a logging job the FMI constantly adjusts the level of the current injected into the formation to maintain a reasonable signal response and signal-to-noise (S/N) ratio, regardless of the formation resistivity. Therefore, a correction must be done to account for the variations in the focusing current applied during logging. Also, the EMEX correction is needed to make the FMI output “comparable to a calibrated resistivity tool” (Grace and Newberry, 1998).

47

3.3.4.2. Data Equalization: Equalizing is a statistical modification of the data to make the signal of each electrode have the same resistivity response as the rest of the electrodes on the tool (Grace and Newberry, 1998). That is, all curves should have the same sensitivity range (Rider, 1996). BorEID uses a moving-window processing technique. In every interpreter-defined vertical window, data from all electrodes are considered and equalized. Then, the software moves, by an interpreter-defined overlapping step, to the next level equalizing the data until the whole logged interval is covered. This process, if done with the right window length and step interval, can “maintain consistent results over the entire length of the log” (Grace and Newberry, 1998). The result of the equalizing process is to have the mean and standard deviation of the signal from every electrode the same as the mean and standard deviation of all “the data from the whole pad over the entire processed interval” (Grace and Newberry, 1998). The module can indicate faulty electrodes by comparing their output to that of the remaining electrodes on the pad. An average signal (as seen from adjacent electrodes) is used to replace the signal of a “dead’ or faulty electrode. In this way, the response from all pads will be uniform and data from “extreme ends of the histogram” (blurred image due to a bad pad contact, washouts, or noise) are filtered. Only data from the rock matrix are kept and equalized.

3.3.4.3. Depth and Speed Corrections: Having, on one hand, the rows of electrodes 0.3 in (0.75 cm) apart on each pad and, on the another hand, the pad and flap rows 5.7 in (14.5 cm) apart, a depth correction has to be done to align all microresistivity curves to the same depth (Figure 3-7). This depth variation is easy to correct when the tool is moving at a constant logging speed, but it becomes complicated when the tool accelerates and/or decelerates when run in a sticky hole. In this case, speed correction will be needed, using the accelerometer data, before any depth correction. Speed and magnetic declination corrections are performed in BorEID. The speed correction is primarily done by using the accelerometer data that gives the tool speed at any depth. Accordingly, a necessary depth shift is applied to position all electrode measurements to the same depth in the hole. Then, the image data itself is used to finetune the accelerometer speed correction.

48

Figure 3 - 7: A schematic diagram showing depth correction to align the microresistivity curves. Note the importance of having a good speed correction before depth correction. Modified after Rider (1996).

49

3.3.5. BorNor BorNor is used to perform a normalization of the FMI data. The normalization can either be static or dynamic. The static normalization is done choosing a resistivity color scale that represents the resistivity range as seen covering the entire data set or log (Figure 3-8). Dynamic normalization uses a certain color range and scale that represents the resistivity range in a sampling window (Figure 3-8) (Rider, 1996; Grace and Newberry, 1998). The advantage of this process is to enhance the sharpness of the image and the contrast between various features seen on it (Figure 3-9). Dynamic normalization was applied to the four FMI logs of this study, using 42 different colors to scale the resistivity image and a 2-ft (61-cm) sliding window.

3.3.6. BorScale The functions and tasks that can be done with BorScale are the conductivity matching and the depth shift correction.

3.3.6.1. Conductivity Matching: “BorScale” provides a technique to match the conductivity of the FMI electrodes to an SFL (Spherically Focused Log) or any other shallow resistivity measurement, such as a shallow laterolog measurement (LLS, HLLS) or a shallow induction measurement (AHT20). The latter has a 20 in (50.8 cm) depth of investigation from the center of the borehole. Processing generates a single electrical curve for each pad by “vertical filtering and lateral averaging” of the FMI image (Grace and Newberry, 1998). Each of these curves is matched to the SFL curve using the leastsquare fitting technique to obtain a fitting coefficient that will be used in unfiltered FMI image files. The scaled FMI image is presented on the log with a gray-color scale. The similarity or difference that the scaled, calibrated image shows compared to an uncalibrated image depends on the resistivity range encountered in the FMI interval being calibrated (Grace and Newberry, 1998).

50

Figure 3 - 8: Image normalization can be done statically by using the resistivity range of the whole logging interval or dynamically by considering formation resistivity ranges in a sliding window. Modified after Rider (1996).

51

Figure 3 - 9: A comparison between a static image (left) and a dynamic image (right). The dynamic image was processed using a 5-ft (1.5-m) sliding window. Note the better appearance of the finely laminated beds and the truncation (T) surface. North (N) on the images refers to true north. Also, the image shows one bedding plane being picked. After Knight (1999) in Hurley (2004).

52

3.3.6.2. Depth Shift: BorNor can depth shift curves acquired during various logging runs to match the FMI curves’ depths. Usually, the GR curve obtained during the FMI run is compared to the GR curve obtained during other runs. Depth shifting can be done automatically or interactively, with the interpreter aligning the GR curve of any logging run to match the one from the FMI logging run. Once the depth shift is defined and applied, a new group of curves (including all the depth-shifted curves) is created. These are presented on the logs with the abbreviation “LDM” (Log Depth Match) in their names.

3.3.7. BorDip BorDip calculates dip and azimuth of bedding planes seen on an FMI image using three dipmeter measurements: the relative positions of three points on a plane, the tool orientation, and the angle and direction of deviation of the tool. Any planar feature on an image log can be defined by its depth, dip, and azimuth. The dip or dip magnitude is the angle, in a vertical plane, between a horizontal line and the line with the “steepest descent” in the plane of interest. The azimuth is the angle between this same line (with the steepest descent) and the geographic north measured in a horizontal plane (Luthi, 2001) (Figure 3- 10). To define a plane, the tool looks for a plane signature on each of the eight microresistivity curves of each pad as seen in Figure 3-5 for a four-pad tool. The difference in depth of the plane signature on each of the microresistivity curves helps determine the dip of the bed. In Geoframe, the dip computation is done by a cross-correlation of the dipmeter microresistivity curves measured from different pads. The method is a mathematical computation of the dips at regular intervals, regardless of the curves’ character. The software does the cross-correlation by comparing “one interval of one dipmeter curve, g, to an interval of the same length, z, of another curve, f, shifted in depth by a displacement h” (Luthi, 2001). Three parameters have to be set for that purpose. These are: the correlation interval length (z), the step length, which is the depth interval at which the correlation procedure is repeated (Luthi, 2001), and the search angle, which is the “range

53

Figure 3 - 10: A schematic diagram showing the dip and azimuth of a bedding plane. After Luthi (2001).

54

on the depth scale searched for correlation before the computer proceeds to another pair of curves at that interval” (Grace and Newberry, 1998). Choosing those parameters, especially the search angle, depends on the interpretation objectives as well as the structural complexity of the area. In areas where high dip magnitude is seen, the search angle is increased. In a well where fine sedimentary details are needed, the correlation interval is usually reduced (increasing the number of dip calculations) and the search angle is kept low (Luthi, 2001). A correlation coefficient, reflecting the curve similarity in shape, is calculated between each two curves using an equation of the form (Luthi, 2001): C(h) = (1/n) * Σz f(z)g(z+h)

(3-1)

Where: n is the number of samples over the correlation interval C is the correlation coefficient This gives a correlogram in which the coefficient C is plotted against the displacement h of the 2 curves. The value of h where C is largest is taken as the most likely correlation of the two curves (Luthi, 2001), and the first curve is shifted at that point by the step length (usually 25-50% of the correlation interval). The whole process is repeated at the next level, giving another dip answer from the correlation of all curve-pair combinations. The result is “the relative position of correlated points around the borehole and a dip answer” (Grace and Newberry, 1998). Having eight pads in the FMI, at any depth, there are 28 cross-correlations (Figure 3-11) between the different curves given by 2 buttons on each of the pads, although only two cross-correlations are needed to define a plane (Grace and Newberry, 1998; Luthi, 2001). The quality of the dip is usually indicated by the software showing the confidence with which the fit was done. A dipmeter tool acquires data at slow and fast sampling rates. Data acquired at slow sampling rates (borehole deviation and azimuth, tool orientation and speed, calipers, EMEX voltage and current, and cable tension) are the ones used for dip calculation (Luthi, 2001). “First, the acceleration data are used to estimate the downhole tool speed and to position each sample of the dipmeter curves at the most likely depth. Then, the

55

Figure 3 - 11: The 28 possible cross-correlations between the FMI pads when determining the dip at a certain depth. After Grace and Newberry (1998).

56

inclinometry data is used to convert apparent dips into true dips knowing the earth geographic coordinates at the well location. The inclinometry data includes the hole deviation and azimuth as well as the relative orientation of the dipmeter sonde with respect to north or the borehole axis” (Luthi, 2001).

3.3.8. BorView “BorView” is the Geoframe module in which the interpreter can interact with the FMI image after all the processing, speed, depth, EMEX corrections, and dip computations have been done. Using mainly the dynamic image/presentation, bedding planes and different types of fractures are traced/picked on the image. When we pick planar features on the image, we trace them on an unrolled image. Therefore, the planar feature appears as a sinusoid or a sine wave. The magnitude of the sine wave is proportional to the dip magnitude and the direction of its trough is the direction of the dip azimuth (Figure 3-12) in a straight hole (Grace and Newberry, 1998). Picking sedimentary features is slightly different from picking fractures. Picking a bedding plane can be done either by choosing at least three points on the plane and letting the computer decide on the best-fit sine wave or by choosing a preset, but interactively adjustable, sine wave and place it on the plane. Fractures are usually picked in a way to have the picked points or traces drawn covering all of the fracture trace seen on the image. Each feature picked on the image is classified in a dip set that describes the feature itself or what it represents. For instance, when picking a fracture, the fracture can belong to either an open natural fracture dip set, a drilling-induced fracture dip set, or a healed fracture dip set. Similarly, when a sedimentologic feature is picked, it is classified as a bedding plane, a scour surface, or a slump. Each dip set is assigned a tadpole shape (circle, diamond, bow-tie, hexagon, triangle, strike, pentagon, and plus sign) and a color to be easily identified on the log. In BorView, also, all the dip and azimuth measurements can be plotted on a stereonet and a rose diagram to calculate a mean dip magnitude and mean dip azimuth for every dip set created.

57

Average depth of fracture = (A+B)/2 Apparent Dip Angle (degrees) = tan-1 {ІA-BІ / D} Where: D is the borehole diameter Figure 3 - 12: A dipping bedding plane in the borehole appearing as a sine wave on the image. The difference in levels of points A and B is used to calculate the apparent dip of the plane. After Serra (1989) in Hurley (2004); modified after Hurley (2004).

58

When we calculate dips directly from images, the software calculate the apparent dip which is the dip in relation to the borehole. In a perfectly vertical borehole, the apparent dip is the same as the true dip. However, every borehole has some deviation, and the dip values represented on the log have to be specified as true dips (TD) or apparent dips (AD) (Grace and Newberry, 1998). The apparent dip is calculated by an equation that relates the height of the sine wave and the diameter of the borehole (Figure 3-12) (Grace and Newberry, 1998): Apparent Dip = tan-1 {ІA-BІ/D}

(3-2)

After plotting the shale beds on a stereonet, the overall structure is revealed as a mean dip magnitude and a mean dip azimuth. Once the “Structural Dip Removal” (SDR) is determined, BorView can recompute all features that have been picked on the image and present them as rotated dips. Inside the BorView module, the structure of all shale beds in the well can be used to derive dips away from the borehole itself in cross sections of arbitrary directions. Many models can be used building the cross-section in order to reveal any subsurface structure such as folding or displacement along a fault plane. Some of the models implemented in StrucView are the monocline model, the similar fold model, the parallel fold model, the concentric fold model, and a rollover model (proprietary of Schlumberger). Figures 3-13 and 3-14 show the shale beds in well Bonanza 4-6 plotted using the monocline and the similar fold models. Note the greater bed continuity and structural details seen using the similar fold model compared to the monocline model.

3.3.9. Data Save Data Save is a module used to save all output files as well as dip sets produced while processing. The module allows us to save the whole Geoframe session (the whole processing chain) in one DLIS file. This one file can be saved on a data tape and can be easily restored.

59

Figure 3 - 13: A structural cross section of the shale beds in well Bonanza 4-6 built using the monocline model. The cross-section direction is 128º from north. (Geoframe software; courtesy of Schlumberger).

60

Figure 3 - 14: A structural cross section of the shale beds in well Bonanza 4-6 built using the similar fold model. The cross-section direction is 260º from north. (Geoframe software; courtesy of Schlumberger).

61

3.3.10. Dip to ASCII The Dip-to-ASCII module helps save a chosen file in ASCII (American Standard Code Information Interchange), LAS (Log ASCII Standard), or dat (data) formats.

3.4. Log Quality Log quality was fair to good in our four wells. The image logs were clear for feature picking except in intervals where the image is masked by gas entry and/or the tool moved faster than the recommended logging speed.

3.4.1. Gas Entry Gas in formation pores may leak into the borehole after drilling. This gas entry forms a thin layer between the tool pad and the borehole wall. The layer of gas masks the borehole wall, and the FMI pad measures the resistivity of the gas itself instead of the formation. As a result, the image in those intervals is blurred, unclear, and, most of the time, featureless (Figure 3-15). Once entered at a deeper level, the gas can be dragged by the pads themselves to shallower depths that may not contain, or be leaking, gas. This deteriorates the image quality even more.

3.4.2. Tool Pull Facing sticky lithology (mainly in mud-swollen shale sections), the FMI pads get stuck for a short duration before being released by the cable pull. Getting freed, the tool may snap upward at a speed higher than the recommended logging speed. Acquisition, in that case, will be beyond the tool sampling rate capabilities. Consequently, not enough data are acquired to show any features on the image (Figure 3-16).

62

Figure 3 - 15: An image log interval in well Bonanza 4-6 showing gas entry from different spots on the borehole wall. Depth scale is feet. Vertical scale is 1:48.

63

Figure 3 - 16: An image log interval from well Pawwinnee 3-181 showing tool pull in separate sections. Note the tool rotation at 10,852 ft. Depth scale is in feet.

64

3.5. Fractures and Structural Analysis Fractures are surfaces of brittle mechanical failure in a rock. They are openings caused by stress release and/or tectonic forces. According to Luthi (2001), fractures can be distinguished based on their morphology. They can be:

1- Irregular fractures associated with brecciation (fault zones, deposition) or formed as a result of collapse after karstification. 2- Solution-enlarged fractures that ultimately become vuggy or cavernous pores. 3- Thin localized joints terminated within the borehole and caused by tectonism, thermal stress, or borehole-related stress amplification. 4- Single, usually planar fractures caused by shearing or tension tectonic movements. 5- Drilling-induced fractures. 6- Healed (resistive) fractures.

Only the last three categories were frequently encountered in the four wells studied, and they will be discussed in details. Grace and Newberry (1998) gave another fracture classification, still based on fracture morphology. According to them, a fracture can be open (filled with drilling mud), healed (filled with minerals), partially healed (some sections are filled with minerals, whereas others are conductive), or vuggy (Figure 3-17). Three factors that affect the appearance of a fracture are Rm (mud resistivity), Rxo (resistivity of the invaded zone), and fracture geometry. Also, according to the same authors, a fracture is either vertical or polygonal. A vertical fracture is a fracture with dip angle of more than 75º, whereas a polygonal fracture has a “chicken wire” appearance caused by dewatering of carbonate or by tectonics (Grace and Newberry, 1998).

3.5.1. Open Fractures Open fractures are easy to see on image logs because, upon drilling, the drilling fluid/mud usually penetrates the fracture opening and creates a thin and conductive sheet 65

Figure 3 - 17: Fracture morphologic types. After Grace and Newberry (1998).

66

with a resistivity greatly different from the resistivity of the rock matrix around it. The current lines flowing through this conductive sheet get modified and become different from the ones flowing in a medium with uniform conductivity, like a rock matrix (Luthi, 2001). On an FMI image, the fractures are seen as features with high dip magnitudes, dark in color (indicating high conductivity), and cutting bedding planes. Open fractures encountered can be of several types: open natural fractures, drilling-induced fractures, and borehole breakouts.

3.5.1.1. Open Natural Fractures: These are fractures present in the formation due to the stress regime in the area and/or tectonic activities. They are independent of the drilling process and are not caused by it. Generally, these fractures are vertical (dip exceeding 75º) and planar in shape (Figure 3-18). Many of the natural open fractures encountered in the studied wells were lithologically bound terminating at bed boundaries and/or when the lithology changed. Some, though, were continuous (Figure 3-18). Open natural fractures for each well were grouped into two main fracture sets: a “continuous fracture set” and a “lithologically bound fracture set”. A “partially healed fracture” set was created in some of the wells where partially healed fractures were encountered. The continuous fractures were traced with red sine waves and represented by red tadpole squares or diamond shapes. The lithologically bound and partially healed fractures were traced with colored traces, shown with blue sine waves on the dynamic image track, and represented with triangular blue tadpoles. Although differentiated when classified, the components of all open natural fractures were combined to compute fracture trace length and fracture porosity in each well. Also, both components were plotted on the same stereonet to compare their strike to the principal stress directions in the well. According to Babcock (1978), the orientation of existing open natural fractures can help predict the strike orientation of induced fractures (discussed below).

3.5.1.2. Drilling-Induced Fractures: They are open fractures caused by the drilling process. Upon drilling, the stress regime is modified due to the removal of rock material from the borehole. “The principal far-field stress gets concentrated as tangential stresses

67

Figure 3 - 18: A section of a dynamic image from well Bonanza 4-6. A continuous open fracture is seen cutting through cross stratifications traced with yellow sine waves and a scour surface traced with a red sine wave. Drilling induced fractures are seen on top of the section.

68

in the two sections of the hole where the wall is parallel to the maximum stress direction” (Luthi, 2001). Orthogonal to those directions, the tangential stress is substantially reduced. The larger the difference in the far-field stress magnitudes, the stronger these effects will be (Luthi, 2001). On the FMI of each of the studied wells, the drillinginduced fractures were grouped into two fracture sets coded by a green color. The sets are a “Drilling-Induced Tensile Fractures” set and a “Drilling-Induced Shear Fractures” set. The symbols assigned to the drilling-induced fracture sets are green “strike” and “bowtie” symbols, respectively. The criterion followed to differentiate between tensile and shear induced fractures is the occurrence of the drilling-induced fractures and their geometry. When occurring vertically, whether straight or branching, those fractures are classified as drilling-induced tensile fractures. When they occur en-echelon, curving at their tips, and, at an angle from the vertical direction, they were classified as shear drilling-induced fractures. The way to differentiate between a natural fracture and a drilling-induced one, when they happen to have the same strike direction, is by studying their shape and occurrence. Induced fractures can be recognized by their irregular shape, branching or en-echelon occurrence, and discontinuity (Luthi, 2001). They occur “almost exactly vertical and at 180 from each other” (Luthi, 2001) (Figure 3-19), and their strike or azimuth direction is almost always constant in the same borehole, indicating the principal horizontal stress. As a rule of thumb, an induced fracture always shows a geometric relationship to the borehole, while a natural fracture does not. The latter often cuts across the entire borehole without changing its appearance (Luthi, 2001).

3.5.1.3. Borehole Breakouts: Borehole breakouts are defined as “broad and flat curvilinear surfaces which enlarge in the direction of minimum horizontal compression” (Zoback et al., 1985; Plumb and Hickman, 1985) and are the result of the “spalling or caving of the borehole” (Luthi, 2001; Zoback et al., 1985) (Figure 3-20). Borehole breakouts are the result of “localized compressive shear failure” (Zoback et al., 1985) and are not associated with natural fracture traversing the well (Plumb and Hickman, 1985). They are a geometric modification of the borehole and grow until the stress is reduced and fracturing ceases. The pads may not be affected by breakout formation as they follow the curvature of the borehole to always be in contact with the borehole wall. 69

Figure 3 - 19: An FMI image showing vertical drilling-induced fractures. Such fractures were classified as drilling-induced tensile fractures on the studied well images. Depth scale is in feet. Modified after Luthi (2001).

70

Figure 3 - 20: A section of a dynamic FMI image from well Bonanza 4-6 showing borehole breakouts and a drilling-induced fracture at 90º from each other. Also, a minor gas entry is seen masking one side of the borehole where the drillinginduced fracture is expected.

71

The caliper, on the other hand, can be indicative of their presence. Borehole breakouts are associated with the slowing or cessation of tool rotation (Babcok, 1978), and can be avoided by increasing the well bore pressure by increasing the mud weight adding additives (like barite) to it (Zoback et al., 1985). Their development is proportional to the difference between the horizontal stress magnitudes, and they occur 90º from both tensile and shear drilling-induced fractures. In other words, the borehole witnesses minimum spalling when “the two effective horizontal stresses are about equal” (Zoback et al., 1985). The smallest borehole breakout that can be detected is related to the borehole diameter itself and the dimension of the tool pad (Plumb and Hickman, 1985).

3.5.1.4. Borehole Elongations: Borehole elongations are borehole failures that can develop in any direction. They were identified by Plumb and Hickman (1985), who had no explanation for their origin. The authors explained that these elongations, sometime developing at 90º to borehole breakout, cannot be natural fractures as “the amplitude of the conductivity anomalies are much smaller than anomalies interpreted as natural fractures.” The authors called them asymmetric borehole elongations (Figure 3-21; Plumb and Hickman, 1985), as the minor gas entry is seen masking one side of the borehole where the drilling-induced fracture is expected. Asymmetry is the sole feature that differentiates them from borehole breakouts. Borehole elongations were rarely seen in the studied wells, and they were not picked nor grouped in dip sets. Borehole breakouts were sufficient to determine the minimum horizontal stress directions in each well. It has been established and demonstrated above that the drilling-induced fractures as well as the borehole breakouts will indicate the direction of maximum and minimum principal horizontal stresses, respectively. Also, it will be expected to have the strike of the open natural fractures trending parallel to the direction of the maximum stress. Therefore, the relationship of all fractures to each other can be summarized as shown in Figure 3-22.

72

(B)

(A)

Figure 3 - 21: An example of an asymmetric elongation in well section (A) with and (B) without borehole breakouts forming in the well. Modified after Plumb and Hickman (1985).

73

Figure 3 - 22: A diagram showing the main fracture classifications in a well and their orientation with respect to the maximum and minimum stress directions in the field. (Courtesy of Jim Urdea, DCS – Schlumberger).

74

3.5.2. Healed Fractures The previous discussion on mud entering fractures is valid for open fractures (natural, drilling-induced, or breakouts). Healed fractures are not open to allow mud entry. They are filled with mineral cement. When a pad is facing one of these fractures, the “current lines get squeezed”, giving rise to an artificially high resistivity measurement. As soon as the tool passes the fracture, “the current lines diverge stronger than they would” (without the presence of a cemented fracture), reflecting an apparent low resistivity. This is how a healed fracture, although thinner than the tool resolution, can be detected. The tool measures the change in resistivity from one side of the fracture to the other, which creates the “halo-like” effect (Figure 3-23) seen on the image and indicative of a healed/resistive fracture (Luthi, 2001).

3.5.3. Fracture Aperture Although a fracture can be smaller than the electrode diameter, it can still be seen on an FMI log because of the great contrast in resistivity between the matrix material and mud found in the fracture. This causes an increase in the electrode current in front of a thin fracture (Figure 3-24). The inversion of this increase in current into the fracture aperture becomes possible from the relationship (Luthi, 2001; Grace and Newberry, 1998): W = a * A * Rmb * Rxo1-b

(3-3)

Where: W = fracture aperture A = the sum of the extra current divided by voltage and integrated along a line perpendicular to the fracture trace (Grace and Newberry, 1998) a = a tool constant b = a value slightly smaller than one Rm = mud resistivity Rxo = resistivity of invaded zone Fractures can be modified in the vicinity of a borehole due to the drilling process and this may increase their apertures. The calculation of the fracture aperture is affected

75

Figure 3 - 23: Two healed fractures are seen with the halo effect (bright areas) seen around them. The upper fracture is totally healed whereas the lower one may be a partially healed fracture. After Luthi (2001) and Grace and Newberry (1998).

76

Figure 3 - 24: (A) Current lines coming out from a pad in front of a fracture calculated by a finite-element modeling code, and (B) response of an FMI electrode when passing a conductive fracture as a function of standoff (s.o.). Note how the current increases even before the electrode is in physical contact with a fracture. The areas under all curves are the same, regardless of the standoff. From Luthi and Souhaite (1990) in Luthi (2001).

77

by Rm. The more conductive the mud is, the clearer the fracture is seen on the image. Also, the way the image is scaled when doing the static normalization is another factor affecting aperture calculation because the software looks at the matrix resistivity around a fracture and the contrast in resistivity between the matrix resistivity and fracture resistivity. The fracture aperture is calculated as either the mean aperture or the hydraulic aperture. The former is the average of the fracture width along its length, whereas the latter is the cubic mean of the fracture width (Grace and Newberry, 1998). The aperture presented on the logs was calculated as the mean aperture. Fracture aperture for drilling-induced fractures are not calculated. They are not meaningful because the fractures are not naturally occurring, in addition to the fact that the fractures are “tapering out toward” their limits (Luthi, 2001).

3.5.4. Fracture Porosity Fracture porosity in each well was calculated for open natural fractures (continuous, lithologically bound, and partially healed). Fracture porosity is “the percentage of the borehole wall (aperture x trace length) covered by fractures to the total borehole volume, which equals the total borehole wall area multiplied by the borehole total depth (Randy Koepsell; personal communication). For the fracture porosity calculation, three parameters are needed: 1) the FMI borehole coverage is measured as seen from the image. It is the ratio of width of the pad-track widths (areas covered by the pad) to the total width (tracks and gap areas) of the image, 2) the fracture aperture is calculated by equation 3-2, and 3) the trace length is specified and picked by the interpreter. Fracture porosity includes the fracture void space and not the matrix porosity. It is presented on the log as a value curve, with the increment seen as added by each fracture.

78

3.5.5. Fracture Total Trace Length Fracture trace length is the segment length covering the trace of a fracture seen on the image. Fracture trace length is added to the log with the increment seen as added by each fracture. Fracture density is shown on the log as well as the number of fractures per foot of borehole along a line perpendicular to the fracture plane.

3.6. Bedding planes and Sedimentary Analysis Bedding planes were traced and picked either by choosing at least three points on the plane and letting the computer decide on the best-fit sine wave, or by choosing an interactively adjustable sine wave and placing it on the plane (as mentioned before). The bedding planes were traced in both sandstones and shaly formations. The bedding plane traces were divided into two different sets. A “Sedimentary_Dip” set includes all bedding planes picked in clean formations, and a “Bed_Boundary_Dip” set contains all picks in shale and shaly sandstone beds. The main objective behind picking bedding planes is to differentiate the sand from the shale and divide the sand layers into different facies in order to reveal the environment in which they were deposited. In order to do that, the FMI images, the dip patterns, and other open-hole logs were examined.

3.6.1. Lithology Determination Lithology was determined using the quicklook lithology column that is built using algorithms involving all open-hole measurements (density, porosity, photoelectric factor, gamma ray, and resistivity) measured by the Schlumberger’s Platform Express (PEX) tool.

79

3.6.2. Environments of Deposition Two main environments of deposition existed during the Upper Cretaceous when the Mesaverde Group was being deposited in the Greater Natural Buttes field. Sandstones from both a fluvial environment (Figure 2-7) and a shoreface environment (Figure 2-8) were seen in the studied wells. Each sand package on the log will be studied to be classified as fluvial sand (channel sand or crevasse splay) or shoreface sand.

3.6.3. GR Signature of Sandstone Facies Certain criteria were set and followed for sand facies and environment of deposition interpretations. Works by Shawa (1979), Galloway and Hobday (1996), Rider (1996), and Scholle and Spearing (1982) were reviewed for GR-log signatures in fluvial and shoreface sands. Personal communications were conducted with the geologists of the Schlumberger DCS (Data and Consulting Services) center in Denver to review their methods and criteria for environment interpretation from dip-pattern studies and dipazimuth vector plots (discussed in a later section), as well as open-hole logs. The pattern of the GR curve (Figure 3-25) was used as an indicator of the environment of deposition based on the work of the above mentioned authors. First, the GR log was used to differentiate sand from shale beds. The contact between the mentioned lithologies was one criterion used to differentiate marine from fluvial sand. On one hand, as stated by Shawa (1979), “prime requisite[s] to assigning a sandstone body to a nearshore origin are: gradational lower contact, an upward-coarsening, and an upwardfining change in sedimentary structures (Figures 3-26 and 3-27) from lower flow regime (ripple cross-laminations and cross-bedding often destroyed by burrowing) to an upper flow regime (horizontal bedding). The complete series of shoreface sand (lower, middle, and upper) was not seen on the image logs. On the other hand, a sharp contact with an erosional base indicates the basal contact of a fluvial channel (Figure 3-27). Coarseningfollowed by fining-upward sequences, as well as blocky GR, are all signatures of fluvial sands. Galloway and Hobday (1996) presented general examples of the GR curve shape in the environments of interest. They showed that a blocky GR curve indicates mainly

80

Figure 3 - 25: Schematic diagram showing the expected GR curve signature in different parts of fluvial systems. Modified from Galloway and Hobday (1996).

81

Figure 3 - 26: Tadpole patterns and GR signatures for continental shelf: tide, wave, and current dominated environments. After Gilreath (1987).

82

83

Figure 3 - 27: Tadpole analysis and GR signatures of (A) non marine/continental environment and (B) continental shelf-delta wavedominated environment. Modified after Gilreath (1987).

braided-stream facies (longitudinal transverse bars) (Figure 3-28). Some blocky short GR sections were interpreted as thin crevasse splays when sandwiched between shale beds. However, crevasse splay, as demonstrated by Galloway and Hobday (1996) and Gilreath (1987) have a slight coarsening-upward sequence (Figures 3-25 and 3-27) and look similar to thin shoreface sand interval. Some sand sections were interpreted as amalgamated crevasse splays, as will be discussed in the sections pertaining to the wells where this was seen. To summarize, in a meandering-stream system, point bars are indicated by a sharp basal contact and a fining-upward sequence (Figure 3-29), as fine sediment content increases going up in a point-bar succession. The GR signatures of both shoreface and crevasse splay sands are expected to coarsen upward. A blocky GR curve can still indicate sand-filled channel (braided-stream deposits) and crevasse splays.

3.6.4. Open-Hole Logs Other wireline curves such as the density and neutron porosity curves and their separation helped indicate the porosity in the sand. This helped differentiate crevassesplay sands because porosities in those facies are poorer compared to point-bar sands.

3.6.5. Coal Coal beds, whenever seen, were a good indicator of an upper shoreface, swampy lagoonal, or transitional environment. Coal seams were very rare and did not exceed 2-4 ft (0.61-1.22 m) in thickness. High resistivity and neutron porosity values are expected in coal layers.

3.6.6. FMI Signatures of Sandstone Lithologies The main features that were looked at on the FMI image to determine the environment of depositions are the following:

84

Figure 3 - 28: The GR curve signature of the braided-channel facies (A and B) of Figure 2-9. Modified after Galloway and Hobday (1996).

85

Figure 3 - 29: The GR curve signature of the meandering system bars seen in Figure 210. Modified after Galloway and Hobday (1996).

86

3.6.6.1. Dip Pattern: An attempt to discern depositional environment from thinning and thickening dip patterns (Figure 3-30) was tried, although the approach had proven to be not very helpful in the Rocky Mountain region (Randy Koepsell, 2006; personal communication). Studying the logs, red and blue patterns (Figure 3-30) were interpreted. These patterns indicate thinning and thickening sequences, respectively. Also, green and yellow patterns that indicate no change and chaotic dip patterns, respectively, were seen. Rarely was there a sand layer showing only one pattern. In many instances, a mixture of many patterns was seen.

3.6.6.2. Dip Magnitude: Dip magnitude was helpful in some instances to determine the environment of deposition. A low dip magnitude, in a shoreface environment, indicates upper shoreface environment. The dip pattern of lower and middle shoreface (Figure 211) was found to be chaotic in general and a good indication of hummocks and burrowing. The chaotic pattern was not exclusive to lower and middle shoreface environments, as many meandering-river sand packages indicated various accretion and opposed accretion directions. A decreasing-upward dip angle (although unusually seen) indicated meandering-river deposits.

3.6.6.3. Scour Surfaces: Scour or erosional surfaces were traced (in red) and grouped in a separate dip set. These surfaces were helpful, in many instances, to indicate the base of fluvial sand and the thalweg orientation as they dip towards it. Also, their abundance within a short thickness generally reflects a braided rather than a meandering river system.

3.6.6.4. Dip Azimuth Vector Plots: The dip-azimuth vector plot (DAVP) is a plot of the cumulative sine values of dip azimuth (measured in radians) against the cumulative cosine values of dip azimuth (measured in radians) of the tadpoles in a specified interval. The plot is a map view that shows how the dip azimuth of the sedimentary structures changes as a function of depth. Depth intervals are seen as segments of different colors on the plots (Figure 3-31). DAVP were plotted for shale beds (Figure 3-31A) for

87

(A)

(B)

(C)

(D)

Figure 3 - 30: Different dip patterns: (A) unchanging, (B) thinning upward, (C) thickening upward, and (D) chaotic. Modified from Luthi (2001).

Figure 3 - 31: A dip-azimuth vector plot of (A) all shale beds in well NBU-222. Inflection points seen on the plots are compared to those in the cumulative dip plot. (B) DAVP showing sections of opposed accretions (below 7,685 ft) that may indicate a general river flow in an ENE direction. Refer to text for discussion.

88

inflection points and in every sand package (Figure 3-31B) to see general flow trends. This helped indicate, in some of the sandstones, the water flow direction based on opposed accretions expected in point bars. Unimodal dip trend or dip azimuth distribution (usually to the NE, E, and SE) indicated braided-stream deposits flowing in the same directions. Opposed accretions (ideally at 90º to each other but the angle may vary when looking at either the upstream or downstream section of a point bar) in superimposed sandstones indicated deposits of a meandering stream. Also, the plots of some marine sandstones showed segments with subtle dip azimuth or accretions in one direction. Such accretions indicated progradation and retrogradation events (see examples in Chapter 5).

Other FMI features that indicated braided-stream deposits were the following: •

Internal clasts seen in the sandstones indicating water energy higher than the one seen in meandering rivers

•

Unimodal dip direction and/or absence of opposed accretion on the DAVP

•

Tabular cross bedding (as interpreted by Longman and Koepsell, 2005)

•

Thick sandstone packages suggesting amalgamation rather than channel migration

•

No fining-upward grain size reflected by color change on the FMI.

FMI features that indicated crevasse splays were the following: •

Thin (1-8 ft [0.31-2.4 m]) sandstone layers interbedded with shale

•

Chaotic dip pattern, faint lamination, or the absence of internal structures

•

Association with meandering-stream deposits

•

Shaly sandstone seen as conductive sandstone on the image log.

3.6.7. Paleocurrent Directions from Image Logs The stereonet plots of the sedimentary structure/features on the FMI logs were not very helpful to determine one main trend or paleocurrent direction in the different modeled intervals (Chapter 6) of the Mesaverde Group. Similar conclusions were reached by Donselaar and Schmidt (2005), who showed that fluvial-channel trends in outcrop, when compared to the “distribution of dip directions in the borehole images show an

89

apparent discrepancy.” Therefore, the task of differentiating paleocurrent direction from accretion was not an objective by itself in the study. In many instances, channel amalgamation made it difficult to separate the exact number of events or channel-fill episodes. Also, the maximum dip direction was uncertain, even when cross-bed sets were recognized. All of the previous reasons can be added to the wide paleocurrent and lateral accretion direction range or distribution expected in a meandering fluvial system and/or a combination of meandering and braided-river deposits.

3.6.8. Structural Dip Removal Bedding planes in both clean and shaly formations were traced to build different dip sets. The traces and tadpoles that represent each bedding plane were color coded (log presentation discussed later). Mean dip magnitude and dip azimuth were obtained for each dip set by plotting the data on either Wulff or Schmidt stereonet plots. The shale bed mean dip magnitude and azimuth (an example in Figure 3-32) were considered to be equal to the structure to be removed to restore the shales back to their horizontal configuration. It is assumed that they were horizontal during deposition. Plotted on a Schmidt plot (Figure 3-32), the poles of the shale beds can be seen to fall on one or more great circles. The number of shale bed pole clusters (or great circles) is the number of intervals that the well should be divided into for various structural removals. When that is the case, the log is divided into various sections, and the mean dip magnitude and mean azimuth of the shale in each is determined. Consequently, the corresponding structural removal is done for each section. However, having all the shale bed poles falling on one great circle, as shown in well NBU-222 (Figure 3-32), indicates that only one structural removal is needed for the whole well. Once determined, the structural removal is applied to all features picked on the log. Relative dips are recomputed for all the sedimentary and structural elements. New sedimentary and bed-boundary dip sets are created with the acronym SDR (Structural Dip Removal) appearing with their names. Those are the ones used for sand layer dip-azimuth vector plots (discussed later).

90

Figure 3 - 32: A Schmidt plot prepared using Geoframe for a shaly interval showing the mean dip value and dip azimuth of the same interval. In this case, all the shaly beds in well NBU-222 were plotted. Their mean dip and dip azimuth are 2.7 and 324 degrees, respectively.

91

3.6.9. Cumulative Dip Plot A cumulative dip plot (Figure 3-33) is created for the shaly bed picks in each well following the method proposed by Hurley (1994). Gamma ray (GR) values were interpolated to the same depths where bedding planes were picked. Various GR cutoffs (80 or 90 GAPI) were applied to filter clean beds. The remaining data were sorted in an ascending depth order. A sample number was given to each measurement or entry in the data set so that the shallowest measurement has a sample number of 1 (Table 3-4). Therefore, the sample numbers are equivalent to depths. The cumulative dip was calculated in a way to have the dip of each measurement added to the sum of dip values of all previous measurements. A plot of the sample number against the cumulative dip was generated (Figure 3-33). The inflection points seen on the plot are good candidates and indicators of faults, micro-faults, unconformities, or slumps (Figure 3-34). They are useful to draw attention to the depth equivalent to an inflection point’s sample number, as the above mentioned features, especially unconformities, may not be obvious on the log, but their depths should be indicated.

Table 3 - 4: Part of the spreadsheet with the data and the cumulative dip calculations.

92

Figure 3 - 33: Example of a cumulative dip plot of the shale beds (GR cutoff is 90 GAPI) in well NBU-222. The plot depth ranges from 6,920 ft (2,109 m; sample number 1) to 9,575 ft (2,918 m; sample number 1422).

93

Figure 3 - 34: A section of well Pawwinnee 3-181 showing slumps and sedimentary deformation. Depth is in feet.

94

3.7. Log Presentation The four logs studied were saved as PDS (Picture Description System) graphics. One presentation format was adopted that consisted of eight tracks. This section describes components and features shown in each track.

3.7.1. Track One Track one contains the following channels and measurements (Figure 3-35):

1- The FMI caliper readings (C1 and C2), scaled depending on the drill bit diameter. Both curves are represented dotted in blue and red, respectively. 2- The PEX caliper reading (HCAL), scaled according to the drill bit diameter. HCAL is represented as a dashed black curve. 3- Tadpole showing the borehole drift and FMI pad one azimuth directions (PAZ1). The tadpole is an open black circle with two tails. A long tail indicates the direction of the borehole drift, and a short tail indicates the direction of PAZ1. 4- The log scale. 5- Net GR: a value curve indicating the increment of the sand-layer thickness using the GR curve. 6- Net Coal: a value curve indicating the increment of the coal bed thickness in the well.

3.7.2. Track Two Track two (Figure 3-35) contains the following features:

1- The FMI statically scaled image. 2- The GR curves from the FMI log run (noted GR_FMI) and the PEX run (noted GR_LDM). Note the latter was depth shifted to match the FMI GR curve.

95

Figure 3 - 35: The first five tracks of the FMI PDS log presentation.

96

3- Fracture traces of all the fractures traced on the dynamic image (track six). Drillinginduced fractures, borehole breakouts, and healed fractures are shown with green, purple, and blue traces, respectively. Open natural fractures (continuous, lithologically bound, and partially filled) were traced with Multicolor-scaled traces.

3.7.3. Track Three Track three (Figure 3-35) contains the following channels and measurements:

1- The resistivity curves, whether laterolog (HLLS, HLLD) or induction (AHT10, AHT20, AHT30, AHT60, and AHT90) curves with various depths of investigation, scaled from 0.1 to 1,000 ohm-m. 2- An average microresistivity curve (SRES), representing the resistivity measurement of one FMI pad (usually Pad eight), scaled from 0.1 to 1,000 ohm-m. 3- Net Fracture Porosity is represented as a value curve indicating the incremental increase in fracture porosity added by every open natural (continuous, lithologically bound, and partially filled) fracture. 4- Mean Fracture Aperture of open natural (continuous, lithologically bound, and partially filled) fractures are represented and scaled from 0.0001 to 1 in.

3.7.4. Track Four The fourth track on the log presentation (Figure 3-35) contains the FMI GR curve, the log scale, and a lithology column.

97

3.7.5. Track Five The fifth track (Figure 3-35) contains the following channels and measurements:

1- The neutron porosity measurement (NPOR_LDM) and the density porosity measurement (DPHZ_LDM) curves, scaled from 40% to -10%. Note that both porosity curves are depth-shifted to match the FMI GR curve. 2- The borehole breakout and drilling–induced fracture trace lengths are represented in purple and green, respectively, to indicate the log intervals containing drillinginduced fractures and borehole breakouts. Areas between the trace lengths and the track margins are shaded to clearly show both represented fractures. 3- Net fracture trace length is presented as a value curve showing the incremental increase in trace length as added by each open natural (continuous, lithologically bound, and partially filled) fracture.

3.7.6. Track Six The sixth track (Figure 3-36) contains the following log features:

1- The FMI dynamically scaled image. 2- All of the sine waves that trace the bedding planes (grouped in the “Sedimentary_Dip” set”, the “Bed_Boundary_Dip” set, and the “Erosional-Scour Surface_Dip” set). Refer to Table 3-5 for the color coding of various sine waves. 3- All of the sine waves that trace the open natural (continuous, lithologically bound, and partially filled) and healed fractures. Refer to Table 3-5 for the color coding of various sine waves. 4- The cable tension curve recorded during the FMI log run. 5- The traces that indicate the drilling-induced fractures and borehole breakouts.

98

Figure 3 - 36: The sixth track of the FMI log presentation.

99

Table 3 - 5: A summary of the dip sets, their color codes and tadpole symbols.

3.7.7. Track Seven The seventh track (Figure 3-37) contains the following log features:

1- All of the tadpoles that indicate the dip magnitudes and dip azimuths of different features picked on the FMI image. The dip magnitude is scaled from -5 to 95 degrees on the track. The dip azimuth is indicated by the direction of the tail on the tadpole having north pointing vertically upward. Refer to Table 3-5 for the tadpole symbols and colors that represent various dip sets and features on the log. 2- The fracture porosity is represented as a red dotted curve scaled from 0 to 1%. 3- The open natural (continuous, lithologically bound, and partially filled) fractures are represented as a blue solid curve scaled as foot of fracture trace per foot of borehole interval.

100

Figure 3 - 37: Tracks seven and eight with all their contents as represented on the FMI log presentation.

101

3.7.8. Track Eight This track (Figure 3-37) contains a value curve that indicates the number of open natural fractures in the well. Note that the software, by default, does not count the last fracture picked in the well. Therefore, the fracture number seen on top of the curve is the number of total open natural fractures minus one.

102

CHAPTER FOUR

BOREHOLE IMAGES: FRACTURE INTERPRETATION

This chapter is a review of the results obtained studying the structural elements in the four studied wells. The structural features seen were different types of fractures in sand and shale layers (no fractures were picked in coal layers). No faults or micro-faults were traced on the image logs although some were suspected on the cumulative dip plots and the StrucView cross-sections. The fracture types encountered are mainly open natural fractures, drilling-induced fractures, healed fractures, and borehole breakouts. Refer to Chapter 3 for the definition of each fracture type. The fractures were traced, and the software determined the fracture plane and assigned a dip magnitude and dip azimuth for each. Individual fracture attributes are provided in ASCII files provided on the attached CD-ROM. Fractures were grouped in different sets. Each set was studied separately to calculate mean dip magnitude, mean dip azimuth, and mean strike direction. The main objective behind determining the strike directions of fracture sets is to reveal the directions of the principal horizontal stresses in the well. As discussed in Chapter 3, the strike directions of natural open and drilling-induced fractures are parallel to the direction of the maximum horizontal stress. On the other hand, borehole breakouts develop and strike parallel to the minimum horizontal stress. After presenting the structural data for all four wells, a comparison of the principal stress directions in each, as calculated from different fracture sets, is presented.

4.1. Well One: Natural Buttes Unit 222 Well Natural Buttes Unit 222 (NBU-222) is located in Natural Buttes field in Utah, in the SW-NE portion of Section 11 (T10S and R22E) (Figure 1-2). The well was logged on February 1, 2000. The FMI log covered a depth interval ranging from 6,905 ft

103

(2,105 m) to 9,599 ft (2,926 m) of measured depth. This gives a total of 2,694 ft (821 m) of logged interval. In NBU-222, the fractures are classified into four different fracture sets. These are: 1) an open natural fracture set, 2) a healed fracture set, 3) a borehole breakout set, and 4) drilling-induced fracture set (with sub-sets differentiating tensile from shear drilling-induced fractures).

4.1.1. Open Natural Fractures A total of 86 open natural fractures were traced in well NBU-222 (Figure 4-1). In general, these were lithologically bound fractures. Occasionally, continuous fractures were seen. Therefore, only one fracture set was created to group all open natural fractures. Each fracture was assigned a dip magnitude and dip azimuth. Plotted on a stereonet, the dip magnitude and dip azimuth means were obtained for the set. These were 86 and 17 degrees, respectively. The mean strike direction of these fractures is 107 degrees from north (Figure 4-1). Trace lengths, trace apertures (knowing the mud weight and resistivity), and fracture porosities were calculated by Schlumberger’s Geoframe software package based on the aperture calculations published by Luthi and Souhaite (1990) and discussed in Chapter 3. The cumulative trace length for all open fractures is 193.6 ft (59 m). The increment or the single fracture contribution to the calculated cumulative can be seen on the log (Refer to FMI image log on the attached data CD). Fracture aperture ranged between 0.00015 and 0.0016 in (0.00381-0.04064 mm), and the cumulative fracture porosity was found to be 0.84%. The increment or the single fracture contribution to the calculated cumulative fracture porosity can be seen on the log.

104

Figure 4 - 1: A stereonet plot, rose diagram, and dip histogram of the open natural fractures traced in well NBU-222. The mean strike direction is 107 degrees. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

105

4.1.2. Drilling-Induced Fractures Drilling-induced fractures were traced on the image log in well NBU-222, and the software package calculated a dip magnitude and a dip azimuth for every fracture manually traced on the image. The induced fractures were differentiated into drilling-induced shear fractures and drilling-induced tensile ones. A total of 89 traces delineating drilling-induced shear fractures and 285 traces delineating drilling-induced tensile fractures were saved in corresponding fracture sets (Figure 4-2). Plotting the fractures on a stereonet, the mean dip and dip azimuth for each drilling-induced fracture sub-set were obtained. These were found to be 86 degrees and 15 degrees, respectively, for the drilling-induced shear fracture set, and 89 degrees and 16 degrees, respectively, for the drilling-induced tensile fracture set. As seen on the rose diagram (Figure 4-2), the mean strike direction of the drilling-induced fractures is 106 degrees, which is mainly the same as the open natural factures.

4.1.3. Borehole Breakouts Borehole breakouts were traced on the image log of well NBU-222, and the software package calculated a dip magnitude and dip azimuth for every breakout manually traced on the image. The traces were not of equal lengths (as no individual breakout trace length needs to be computed), but covered the total breakout fracture length seen on the image. A total of 67 borehole breakout traces were plotted on a stereonet (Figure 4-3). The stereonet shows a mean dip of 89 degrees and a mean dip azimuth of 108 degrees. The rose diagram of the breakouts (Figure 4-3) shows a slight scatter in the strike value, ranging between north and 35 degrees from north. However, a mean strike direction for the set trends 18 degrees from north, which is about 90º different from open natural/drilling-induced fractures.

106

Figure 4 - 2: A stereonet plot, rose diagram, and dip histogram of the drilling-induced shear and tensile fractures in well NBU-222. The mean strike direction of the drilling-induced fracture set is 106 degrees. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

107

Figure 4 - 3: A stereonet plot, rose diagram, and dip histogram of the borehole breakouts in well NBU-222. The mean strike direction of this set is 18 degrees. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

108

4.1.4. Healed (Resistive) Fractures Only one fracture was traced and interpreted to be a healed fracture in well NBU222. The fracture is at 7,312.24 ft (2,228.77 m) depth. It has a dip magnitude of 89 degrees and a dip azimuth of 352 degrees. The strike of this fracture is, thus, 82 degrees from north.

4.2. Well Two: Bonanza 4-6 Well Bonanza 4-6 is located in the Bonanza field in Utah, SE-NE-SW portion of Section four (T10S and R23E) (Figure 1-2). The well was logged on September 7, 2003. The FMI log covered a depth interval ranging from 6,800 ft (2,073 m) to 8,510 ft (2,594 m) of measured depth. This gives a total of 1,710 ft (521 m) of logged interval. In Bonanza 4-6, the fractures are classified into five different fracture sets. These are: 1) an open natural fracture set, 2) a partially healed fracture set, 3) a resistive fracture set, 4) a borehole breakout set, and 5) a drilling-induced fracture set (with sub-sets differentiating tensile from shear drilling-induced fractures).

4.2.1. Open Natural and Partially Healed Fractures A total of 38 open natural fractures and 23 partially healed fractures were traced in well Bonanza 4-6. The open fractures were mainly lithologically bound fractures except for two that were continuous. Each fracture was assigned a dip magnitude and dip azimuth. Plotted on a stereonet, dip magnitude and dip azimuth means were obtained for each of the fracture types mentioned above (open natural and the partially healed). These were 89 degrees of dip magnitude and 24 degrees of dip azimuth for the open natural fracture set, and 86 degrees of dip magnitude and 7 degrees of dip azimuth for the partially healed fracture set (Figure 4-4). Due to gas entry in the well (refer to the FMI image log saved as a PDS file on the attached data CD-ROM), the image was of poor quality. In many instances, tracing the

109

Figure 4 - 4: A stereonet plot, rose diagram, and dip histogram of the open natural and partially healed fractures traced in well Bonanza 4-6. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

110

partially healed fractures was very interpretive following a gas entry pattern that fit a sinusoid, and this caused a wide range of fracture strike trends. Therefore, for a mean dip magnitude and mean dip azimuth for this fracture set, only the open fractures will be considered. The mean strike direction for the open fractures in Bonanza 4-6 is 114 degrees. The cumulative trace length for all open fractures in the well is 81 ft (24.8 m). The increment or the single fracture contribution to the calculated cumulative can be seen on the log (refer to FMI image log on the attached data CD-ROM). Fracture aperture ranged between 0.00075 and 0.017 in (0.01905-0.4318 mm), and the cumulative fracture porosity was found to be 4.51%. The incremental single fracture contribution to the calculated cumulative fracture porosity can be seen on the log. The fracture porosity in this well is an order of magnitude higher than the other three wells. This optimistic value may not be realistic for a total of 61 traced open and partially healed fractures and may be due to a subjective FMI image scaling that affected the aperture computation.

4.2.2. Drilling-Induced Fractures Drilling-induced fractures were traced on the image log of well Bonanza 4-6, and the software package calculated a dip magnitude and dip azimuth for every fracture manually traced on the image. The traces were not of equal length (as no individual induced fracture trace length needs to be computed), but covered the total induced fracture length seen on the image. The drilling-induced fractures were differentiated into shear and tensile fractures. A total of 66 traces delineating drilling-induced shear fractures and 98 traces delineating drilling-induced tensile fractures were saved in each of the corresponding fracture sets (Figure 4-5). Plotting the drilling-induced fractures on a stereonet, the mean dip magnitude and dip azimuth are 88 degrees and 10 degrees, respectively, for the drilling-induced shear fracture set. The mean dip magnitude and dip azimuth are 89 degrees and 191 degrees, respectively, for the drilling-induced tensile fracture set. As seen on the rose diagram (Figure 4-5), the mean strike direction of the drilling-induced fractures is 100 degrees.

111

Figure 4 - 5: A stereonet plot, rose diagram, and dip histogram of the drilling-induced (shear & tensile) fractures and borehole breakouts in well Bonanza 4-6. The mean strike directions for the drilling-induced fractures and borehole breakouts are 100 and 18 degrees, respectively. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

112

4.2.3. Borehole Breakouts Borehole breakouts were traced on the image log. A total of 114 breakout traces were obtained and plotted on a stereonet (Figure 4-5). The stereonet shows a mean dip of 88 degrees and a mean dip azimuth of 108 degrees. The rose diagram of the breakouts (Figure 4-5) shows a slight scatter in the strike value, ranging between north and 35 degrees from north. A mean strike direction for the set is 18 degrees.

4.2.4. Healed (Resistive) Fractures Three fractures were interpreted as healed fractures in well Bonanza 4-6. Each interpreted healed fracture had a dip magnitude and a dip azimuth different from the other two (Figure 4-6). Neither a mean dip magnitude nor a dip azimuth is calculated, and the paleo-stress direction could not be inferred from the fractures.

4.3. Well Three: Pawwinnee 3-181 Well Pawwinnee 3-181 is located in the Pawwinnee field in Utah, NW-NW-NW portion of Section 3, (T9S and R21E) (Figure 1-2). The well was logged on March 27, 2003. Two logging runs were conducted to cover the interval of interest extending from 9,468 ft (2,886 m) to 12,222 ft (3,725 m) of measured depth. The total logged interval was 2,754 ft (839 m), divided in two raw FMI data files with 229 ft (70 m) of overlap between both runs. Each of the files was processed independently in Geoframe, and the two intervals will be referred to as the deeper well interval and the shallower well interval. Therefore, two stereonet plots for each fracture set will be presented to cover the structural analysis of both intervals. The fractures in the well are classified into four different sets. These are: 1) a lithologically bound fracture set, 2) a partially healed fracture set, 3) a healed fracture set, 4) a borehole breakout set, and 5) a drilling-induced tensile fracture set (only three drilling-induced fractures were interpreted as shear fractures).

113

Figure 4 - 6: A stereonet plot, rose diagram, and dip histogram of the healed fractures in well Bonanza 4-6. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

114

4.3.1. Open Natural and Partially Healed Fractures In the shallower well interval, a total of 13 lithologically bound open fractures and four partially healed fractures (Figure 4-7) were traced. Each fracture was assigned a dip magnitude and a dip azimuth. Plotted on a stereonet, the dip magnitude and dip azimuth means of the lithologically bound fractures were found to be 82 and 45 degrees, respectively. The dip magnitude and dip azimuth means of the partially healed fractures were found to be 23 and 319 degrees, respectively. In the deeper well interval, a total of 15 lithologically bound fractures were traced (Figure 4-8). Each fracture was assigned a dip magnitude and dip azimuth. Plotted on a stereonet, the dip magnitude and dip azimuth means were found to be 88 and 12 degrees, respectively. Seven of the lithologically bound fractures in the shallower interval (Figure 4-7) show a scatter in their dip magnitudes as well as their dip azimuths, affecting the mean strike direction calculated for the fracture set. This is giving a difference of 33 degrees in the calculated mean strike direction of the open fractures in both intervals (135 degrees in the shallower interval vs. 102 degrees in deeper interval). Those fractures will be ignored, and the mean strike direction of the deeper interval’s fractures (102 degrees) will be considered to be the strike direction of the open natural fracture in the well. The cumulative trace length for all open fractures (in both intervals) is 57.2 ft (17.4 m). The increment or the single fracture contribution to the calculated cumulative can be seen on the log (refer to FMI image log on the attached data CD-ROM). Fracture aperture, considering only the open natural fractures in both intervals, ranged between 0.0017 and 0.015 in (0.4318-0.381 mm). The cumulative fracture porosity was found to be 0.19%. The incremental single fracture contribution to the calculated cumulative fracture porosity can be seen on the log (refer to FMI image log on the attached data CD-ROM).

115

Figure 4 - 7: A stereonet plot, rose diagram, and dip histogram of the open natural and partially healed fractures traced in the shallower interval of well Pawwinnee 3181. The mean strike direction of the natural open fractures is 135 degrees. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

116

Figure 4 - 8: A stereonet plot, rose diagram, and dip histogram of the open natural fractures traced in the deeper interval in well Pawwinnee 3-181. The mean strike direction of the fracture set is 102 degrees. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

117

4.3.2. Drilling-Induced Fractures Drilling-induced fractures were traced on the image log. The drilling-induced fractures discussed are the tensile fractures as only three fractures were classified as shear fractures. A total of 49 fractures were traced in the shallower well interval (Figure 4-9) and 16 fractures in the deeper well interval (Figure 4-10). Plotting the drilling-induced fractures on a stereonet, the mean dip magnitude and mean dip azimuth were found to be 88 and 197 degrees, respectively, in the shallower interval and 89 and 191 degrees, respectively, in the deeper interval. This gives a drillinginduced strike direction ranging between 101 and 107 degrees, as seen on the rose diagrams of Figures 4-9 and 4-10. Taking the average of both numbers, the mean strike direction for the drilling-induced fractures in the well becomes 104 degrees.

4.3.3. Borehole Breakouts Borehole breakouts were traced in well Pawwinnee 3-181 on the image log, and the software package calculated a dip magnitude and a dip azimuth for every breakout manually traced on the image. A total of 26 breakout traces were obtained from both well intervals. Plotted on a stereonet (Figures 4-9 and 4-10), the shallower interval’s breakouts had a mean dip magnitude and a mean dip azimuth of 88 and 101 degrees, respectively. In the deeper interval, the borehole breakouts had a mean dip magnitude and a mean dip azimuth of 88 and 105 degrees, respectively. The rose diagrams of the breakout fractures (Figures 4-11 and 4-12) show breakout strike directions that range between 11 and 15 degrees. Taking the average of both numbers, the mean strike direction of the borehole breakouts in the well becomes 13 degrees.

118

Figure 4 - 9: A stereonet plot, rose diagram, and dip histogram of the drilling-induced fractures and borehole breakouts in the shallower interval in well Pawwinnee 3181. The mean strike directions for the drilling-induced fractures and borehole breakouts are 107 and 11 degrees, respectively. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

119

Figure 4 - 10: A stereonet plot, rose diagram, and dip histogram of the drilling-induced fractures and borehole breakouts in the deeper interval in well Pawwinnee 3-181. The mean strike directions for the drilling-induced fractures and borehole breakouts are 101 and 15 degrees, respectively. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

120

Figure 4 - 11: A stereonet plot, rose diagram, and dip histogram of the healed fractures picked in the shallower interval of well Pawwinnee 3-181. The mean fracture strike direction is 45 degrees. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

121

Figure 4 - 12: A stereonet plot, rose diagram, and dip histogram of the healed fractures in the deeper interval of well Pawwinnee 3-181. The mean fracture strike direction is 132 degrees. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

122

4.3.4. Healed (Resistive) Fractures A total of 13 healed fractures were traced in both of the intervals. Seven fractures in the shallower interval had a mean dip magnitude and a mean dip azimuth of 12 and 135 degrees, respectively (Figure 4-11). Six healed fractures that were traced in the deeper interval showed a wide range of dip magnitudes as well as dip azimuths (Figure 412). The means were 79 and 42 degrees, respectively. As a result, no mean strike direction can be inferred from the data, as the strike trend scattered over a range between 45 and 132 degrees.

4.4. Well Four: Kennedy Wash Federal Unit 16-1 Well Kennedy Federal Unit (KWFU) 16-1 is located in the KWFU field in Utah, in NW-NW portion of Section 16 (T9S and R21E) (Figure 1-2). The well was logged on October 8, 2000. The FMI log covered a depth interval ranging from 8,490 ft (2,588 m) to 10,991 ft (3,350 m) of measured depth. This gives a total of 2,501 ft (762 m) of logged interval. In well KWFU 16-1, the fractures are classified into four different fracture sets. These are: 1) a lithologically bound fracture set, 2) a partially healed fracture set, 3) a continuous fracture set, 4) a borehole breakout set, and 5) a drilling-induced fracture set (with sub-sets differentiating tensile from shear drilling induced fractures). No healed fractures were found in KWFU 16-1.

4.4.1. Open Natural and Partially Healed Fractures Both the partially healed fracture set and the continuous fracture set (Figure 4-13) contained two fractures each. No mean dip magnitude and dip azimuth were calculated for any of them. Therefore, the analysis of the open natural fractures in this well is that of the lithologically bound fractures.

123

Figure 4 - 13: A stereonet plot, rose diagram, and dip histogram of all open natural fractures traced in well KWFU 16-1. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

124

A total of 13 lithologically bound open fractures were traced in well KWFU 16-1. Each fracture was assigned a dip magnitude and a dip azimuth. Plotted on a stereonet, the mean dip magnitude and mean dip azimuth were found to be 75 and 185 degrees, respectively (Figure 4-13). Consequently, the mean fracture strike direction is 95 degrees. Trace lengths, trace apertures (knowing the mud weight and resistivity), and fracture porosities were calculated. The cumulative trace length for all open fractures is 14.9 ft (4.5 m). The incremental single fracture contribution to the calculated cumulative can be seen on the log (refer to FMI image log on the attached data CD-ROM). Fracture apertures of all natural open fractures ranged between 0.00035 and 0.01 in (0.009 and 0.254 mm). The cumulative fracture porosity was found to be 0.54 %. The incremental single fracture contribution to the calculated cumulative fracture porosity can be seen on the log (refer to the FMI image log on the attached data CD-ROM).

4.4.2. Drilling-Induced Fractures Drilling-induced fractures were traced on the image log of well KWFU 16-1 and were differentiated into shear (74 fractures) and tensile (107 fractures) fractures (Figure 4-14). Plotting the drilling-induced fractures on a stereonet, the mean dip magnitude and mean dip azimuth were found to be 81 and 193 degrees, respectively, for the drillinginduced shear fracture set, and 88 and 190 degrees, respectively, for the drilling-induced tensile fracture set. As seen on the rose diagram (Figure 4-14), the mean strike direction of the drilling-induced fracture sets is 101 degrees from north.

4.4.3. Borehole Breakouts A total of 193 breakout traces were plotted on a stereonet (Figure 4-14). The stereonet shows a mean dip magnitude of 87 degrees and a mean dip azimuth of 108 degrees. The rose diagram of the breakouts (Figure 4-14) shows a slight scatter in the strike value, ranging between north and 35 degrees from north. A mean strike direction is

125

Figure 4 - 14: A stereonet plot, rose diagram, and dip histogram of the drilling-induced (shear and tensile) fractures and borehole breakouts in well KWFU 16-1. The mean strike directions of the drilling-induced fractures and borehole breakouts are 101 and 18 degrees, respectively. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

126

18 degrees from north.

4.5. Summary Table 4-1 summarizes the structural data classified by well. The mean strike direction of each fracture set/type in each well is presented in terms of degrees from north. The results show that the data obtained from both the open natural and drillinginduced fractures are in good agreement, both indicating the direction of the maximum horizontal stress direction (SHmax). It can be concluded that SHmax, taking the average of the data obtained from both mentioned fracture types, trends 103 degrees from north in the GNB field. The borehole breakout strike direction indicates the direction of the minimum horizontal stress (SHmin). The average of the data obtained shows that SHmin trends 17 degrees from north in the field. Both the maximum and minimum horizontal stress directions are nearly 90º from one another; an angle expected to separate the principal horizontal stresses. The healed fractures were seen with a wide range of dip magnitudes and dip azimuths in each well, as well as when comparing each well with the remaining three. No one general trend could be considered as a mean strike direction of the healed fractures in the GNB field. Therefore, no one paleo-stress direction could be inferred. This can be due to the subjectivity with which the healed fractures were traced and interpreted. Also, it can be the result of various stress regimes fracturing the rocks during different times.

4.6. Discussion Fractures in the Mesaverde Group in the Piceance basin have been related to a “westward thrust indentation” (Lorenz, 2003). They are also related to stress and deformation caused by the interaction of the North American and Pacific Plates since the middle Tertiary (Zoback and Zoback, 1989). The average SHmax direction obtained (Table 4-1) matches the WNW-ESE strike direction obtained by Lorenz and Finley (1991)

127

128

Table 4 - 1: Structural analysis summary.

(Figure 4-15A) and referred to by Lorenz (2003) in the Cretaceous Mesaverde in the eastcentral Piceance basin in NW Colorado. Figure 4-15A shows a rose diagram of the strike directions of 62 extension fractures in one well in the Piceance basin, NW Colorado. A study conducted by Koepsell et al. (2003) interpreted FMI images of 23 different wells in the Piceance basin and showed the same strike distribution for open natural fractures (Figure 4-15B) and induced fractures (Figure 4-15C). Considering the stress regimes and stress directions in a broader area in the Rocky Mountain region, the results obtained by Zoback and Zoback (1989) were reviewed. The authors divided the Rocky Mountain region into three stress provinces: Cordillera extensional, Colorado Plateau interior, and the southern Great Plains. They summarized the stress results in the Colorado Plateau (the province that includes the GNB field) to indicate, indeed, a general WNW orientation of SHmax. Therefore, the structural results and stress directions obtained from the four studied wells match observations from both the Mesaverde Group and the Rocky Mountain region. The mean strike directions of open natural and drilling-induced fractures obtained in this study (Table 4-1) and in previous studies done in the Piceance basin (Figure 4-15) do not closely match the Gilsonite dikes direction seen on the geologic maps of Utah (Figure 5-16; Rowley et al., 1985). This may be due to the fact that the orientation of the drilling-induced fractures are contributing more to the calculation of the mean direction as they are more abundant in all the four wells. A closer look at Figures 4-1, 4-4 and 4-7 of wells NBU-222, Bonanza 4-6, and Pawwinnee 3-181 (shallow interval), respectively, shows a wide distribution of the strike directions of the open natural fractures in the wells. This wide range of open natural fracture directions (60º - 170º) is the result of shifts of the stress field with time. Although open natural fractures in various well intervals are not dated or compared to the age of the Gilsonite dikes, they are still older than the drilling-induced fractures. Therefore, we can conclude that the present-day stress regime is more likely to be parallel to the drilling-induced fracture orientations. The importance of knowing the principal stress directions is useful to design hydraulic fracturing jobs in the low permeability sandstones. Hydraulic fractures can significantly increase well productivity in tight reservoirs. Knowledge of fracture direction is required to optimize the results of the fracturing job as hydraulic fractures

129

(A)

(B)

(C)

Figure 4 - 15: Rose diagrams of (A) the 62 vertical extension fractures in the MWX well that shows a main fracture strike in an ESE-WNW direction (Lorenz and Finley, 1991; Lorenz, 2003). (B) The strike direction of open natural fractures, and (C) drilling-induced fractures from the 23 wells interpreted by Koepsell et al. (2003) in the Piceance basin.

130

131

Figure 4 - 16: Section of the geologic map of the Vernal area. Modified after Rowley et al. (1985). Map key is on the following page.

Figure 4-16: (continued from the previous page).

132

develop and propagate parallel to SHmax. Also, field development can be more successful if fracture and fault geometry, distribution, and direction are known. As such, well spacing and location can be better chosen to overcome compartmentalization, design well spacing, and optimize production. Other crucial parameters are the magnitudes of the maximum and minimum stresses. The difference of both magnitudes is useful in designing hydraulic fracturing jobs.

133

CHAPTER FIVE

BOREHOLE IMAGES: FACIES INTERPRETATION

The chapter is a review of the results obtained by studying the sedimentary features (bedding planes, scour surfaces, cross strata, and dip patterns) of the sand packages on the FMI image logs. Taking into consideration the open-hole measurements (GR and PEX) and quicklook lithology determination, the log sections were divided into shaly sections and sand packages. Bedding planes were traced in the four wells on an average of 4 bedding planes per foot in the sand layers and 1 bedding plane per foot in the shaly sections. Geoframe assigned a dip magnitude and dip azimuth to each picked bedding plane. The bedding planes in each well were grouped into dip sets: 1) a “SedimentaryBedding Dip” set that contained all bedding planes and cross strata picked in sand layers, and, 2) a “Bed-Boundary Dip” set that contained all bedding planes picked in shale. Tadpole patterns in each of the sand packages were studied to differentiate sand facies. The sand packages, the cumulative-dip plots (CDP), and the dip-azimuth vector plots (DAVP) will be reviewed for each of the wells. The main objective is to differentiate the sand into various fluvial and shoreface facies, indicate depths where unconformities, slumps, or faults are suspected, attempt to determine the flow and accretion directions in the sand, and present the facies proportions in different zones in each well. Formation tops in the studied wells were obtained from Kerr McGee Corporation (Jerry Cuzella; personal communication, 2006; Table 5-1). The main tops are the top of the Wasatch Formation (WSTCH), the Dark Canyon Formation (DKCYN), the Upper Mesaverde Group (MVU), the Lower Mesaverde Group (MVL or MVL1), the Castlegate Formation (CGATE), and the Mancos Shale (MN or MN1). The wells will be divided into various zones based upon those formation tops.

134

Table 5 - 1: Formation tops in the studied wells as interpreted by the Kerr McGee Corporation. The total depth (TD), Kelly bushing (KB), and formation tops are measured depths reported in feet. Well Name KWFU 16-1 NBU-222 BONANZA 4-6 PAWWINNEE 3-181

KB

TD

5,395 11,009 5,051 9,612 5,357 8,529 4,702 12,241

WSATCH NRHN DKCYN MVU 6,091 4,034 4,182 5,440

7,426 5,486 5,365 7,078

8,432 6,440 6,277 8,603

MVL* MVL1* CGATE

MN1

8,508 9,649 10,033 10,792 6,531 7,839 8,620 8,960 6,352 7,656 8,645 9,944 10,807 11,146

WSATCH = Wasatch Formation, NRHN = North Horn Formation, DKCYN = Dark Canyon Formation, MVU = Upper Mesaverde Group, MVL = Lower Mesaverde Group, CGATE = Castlegate Formation, and MN1 = Mancos Shale. * Kerr McGee divided the Upper and Lower Mesaverde into different (numbered) units without necessarily giving names to various sandstone formations in each. In well NBU-222, MVL1 was considered the top of the Lower Mesaverde Group.

5.1. Well One: Natural Buttes Unit 222 Well Natural Buttes Unit 222 (NBU-222) is located in Natural Buttes field in Utah, in the SW-NE portion of Section 11 (T10S and R22E) (Figure 1-2). The FMI log covered a depth interval ranging from 6,905 ft (2,105 m) to 9,599 ft (2,926 m) of measured depth. A total of 4,587 bedding planes were picked in NBU-222. 2,807 of which were picked in sand beds and grouped in the Sedimentary-Bedding dip set, and 166 bedding planes were interpreted as erosional/scour surfaces (Figure 5-1). 1,614 bedding planes were picked in shale and grouped in the Bed-Boundary dip set (Figure 5-2).

5.1.1. Structural Dip Removal As seen in Figure 5-2, the Bed-Boundary dip set has a mean dip magnitude of 2.17 degrees and a mean dip azimuth of 323 degrees. Assuming that shale beds are horizontal upon deposition, this structure has to be removed to restore the shale beds back to their original horizontal configuration. The structure removed, thus, was 2.17/323, and all the features picked on the FMI image were recomputed and presented in terms of

135

Figure 5 - 1: A stereonet plot, rose diagram, and dip histogram of the “Sedimentary_Dip” set [that contained the bedding planes picked in sand layers (low GR)] in well NBU-222 and the “erosional_scour” dip set [that contains bedding planes interpreted to be erosional surfaces]. The mean dip magnitude and dip azimuth values of both dip sets are the values before the structural dip removal. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

136

Figure 5 - 2: A stereonet plot, rose diagram, and dip histogram of the “Bed_Boundary” dip set [that contained bedding planes picked in shaly (high GR) intervals] in well NBU-222. The mean dip magnitude (2.17 degrees) and dip azimuth (323 degrees) are the values before the structure dip removal. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

137

relative/corrected dip. To create new ASCII files of the recomputed dip sets, the “Dip Removal” module of Geoframe was used. The new sets created are: 1) the “SedimentaryBedding-SDR Dip” set, and 2) the “Bed-Boundary-SDR Dip” set.

5.1.2. Cumulative Dip Plot A CDP (Figure 5-3) was created for the shaly bedding plane picks in the well following the method proposed by Hurley (1994) (Review Chapter Three). Gamma ray (GR) values were interpolated to the same depths where bedding planes were picked, and a 90-GAPI cutoff was applied to the data in well NBU-222 to filter out bedding planes picked in clean formations. A plot of the sample numbers vs. the cumulative dips was plotted (Figure 5-3). This shows inflection points, which can be good indicators of faults, unconformities, or slumps. Minor inflection points on the CDP (Figure 5-3) correspond to the following measured depths: 7,282 ft (2,220 m), 7,358 ft (2,243 m), 7,583 ft (2,311 m), 7,796 ft (2,376 m), 7,846 ft (2,391.5 m), 8,051ft (2,454 m), 8,143 ft (2,482 m), 8704 ft (2,653 m), 8,806 ft (2,684 m), 9,475 ft (2,888 m), and 9,572 ft (2,918 m). The FMI image does not indicate an unconformity, slump, or fault at 7,282 ft (2,220 m) or at the vicinity of that depth [knowing that a fault or slump may occur at ± 10 samples away from an inflection point with an indicated depth]. Fault planes (discussed below) may correspond to some of the inflection points. A tool pull occurred at 9,475 ft (2,888 m), so the FMI image does not reveal any sedimentary structures at that depth. 7,358 ft (2,243 m), 7,796 ft (2,376 m), 7,846 ft (2,391.5 m), 8,051 ft (2,454 m), 8,143 ft (2,482 m), and 8,806 ft (2,684 m) correspond to fining upward (with the exception of 8,051 ft which is the top of a blocky shaly sand package) shaly (GR>90 GAPI), not-filtered, sands covered with shale. The overlying shale has its bedding plane with dip azimuth different than the sand. The sedimentary structure on the image does not indicate slumping at those depths. An erosional surface was traced at 8,704 ft (2,653 m), which corresponds to the base of a channel. Therefore, a (minor) erosional event may be indicated at that depth. A boundary between a shoreface sand package and a thick sequence of marine shale exists at 9,572 ft

138

Figure 5 - 3: A cumulative dip plot of the bedding planes in well NBU-222 after a 90 GAPI cutoff. The arrows point to inflection points and indicate their depths in feet. Data sets in pink, yellow, blue, and cyan represent bedding planes with dip azimuth ranging from 0 to 90, 90 to 180, 180 to 270, and 270 to 360 degrees from north, respectively. The depths of various formation tops are indicated. (MVL = Lower Mesaverde Group, CGATE = Castlegate Formation, MN = Mancos Shale).

139

(2,918 m). This depth may represent the regional unconformity at the base of the CGATE. The CDP (Figure 5-3) shows a major inflection point at sampling point 449 corresponding to a measured depth of 8,165 ft (2,489 m). The top of the well (sample number 0) to this inflection point is an interval with steep dips (Figure 5-4). The dips from the inflection point to the bottom of the well are relatively much shallower (Figures 5-3 and 5-4). This major inflection point is seen at an erosional surface traced on the image log. Above this scour surface, four feet of blocky sand are seen. This significant event may be a fault or an unconformity. Plotting a cross section in StrucView (Figure 5-4) shows a potential fault plane close to the same depth. As mentioned in Chapter Three, StrucView is a structural module in Geoframe that allows us to plot the bedding plane dip set components on a cross section of arbitrary direction using various structural models. In our case, the shale beds were plotted to see their continuity away from the borehole in a cross section trending 125º from north and using the similar-fold model (proprietary of Schlumberger) (Figure 5-4). The cross section shows potential fault planes. These are traced in black. The lowermost fault plane is seen at a depth between 8,100 ft (2,469 m) and 8,150 ft (2,484 m). This may be the feature causing the major inflection point on the cumulative dip plot. Other fault planes are traced at the following measured depths: 7,329 ft (2,234 m), 7,440 ft (2,268 m), 7,652 ft (2,332 m), 7,767 ft (2,367 m), and 8,120 ft (2,475 m). Some of these depths are close to the depths seen on the CDP and may be the cause of some minor inflection points.

5.1.3. Dip Azimuth Vector Plot The DAVP is another plot that may show inflection points at depths where unconformities, slumps, or faults may be present. Figure 5-5 shows the DAVP (built in Geoframe) of all the shale beds, before structural dip removal, in well NBU-222. The plot shows a significant inflection at 8,831 ft (2,692 m). Below 8,831 ft (2,692 m), the plot is straight. This represents the interval with shallow dips as mentioned in the previous section and seen in Figures 5-3 and 5-4. The interval above 8,831 ft (2,692 m) shows 3

140

Figure 5 - 4: A cross section showing the shale bedding planes in well NBU-222. Few planes (traced in black) were interpreted to be potential fault planes. Note the steep dips in the upper 1,400 ft of the well interval. (Data plotted using the “StrucView” module of the Geoframe software package; Courtesy of Schlumberger).

141

Figure 5 - 5: A dip azimuth vector plot of all shale beds in well NBU-222. Shale beds are plotted with no structural dip removal applied. (MVL = Lower Mesaverde Group, CGATE = Castlegate Formation, MN = Mancos Shale).

142

small segments (6,914-7,690 ft, 7,690-8,351 ft, and 8,351-8,831 ft) with different slopes. The inflection point at 8,831 ft (2,692 m) on the DAVP is close to the minor inflection point picked at 8,806 ft (2,684 m) on the CDP. DAVPs, with structurally corrected sand beds, were plotted for every sand package to see general flow trends. This helped indicate, in some of the sands, the flow direction and opposed accretions expected in lateral point bars. The approach proved useful in many cases to differentiate channel fill (longitudinal bars) from lateral point bars.

5.1.4. Facies Examples from Well NBU-222 As mentioned in Chapter Three, two main environments of deposition were encountered in the studied wells. These are the fluvial (braided and meandering stream systems) environment and the shoreface environment. Each sand layer was studied separately to determine its deposition in either environment. The basal and top GR contact, the tadpole patterns, the DAVP of the sand bedding planes, and cross strata were studied to determine sand accretion direction(s) and infer general flow direction(s). Several facies examples from well NBU-222 will be presented before summarizing the facies proportions of the whole well. The remaining sand packages, the FMI image log, and the DAVPs are provided on the attached CDROM. A table with the interpretation of every sand package in the well is provided in Appendix A. The sandstone interval that extends from 7,678 ft (2,340 m) to 7,692 ft (2,345 m) (Figure 5-6) is a section in the upper coaly interval of the Lower Mesaverde (Longman and Koepsell, 2005). The GR shows a blocky signature with sharp sand-shale contact at the base and on top. The sand is interpreted to be a meandering river channel fill. The DAVP of the same interval (Figure 5-7) shows a general sand accretion direction pointing toward the NW between 7,676 ft (2,340 m) and 7,681 ft (2,341 m) and a general accretion towards the S and SE below 7,681 ft (2,341 m). This may have been the result of a river flowing in an easterly direction.

143

144

Figure 5 - 6: Example of a channel sand fill in well NBU-222. Note the blocky GR and the tadpoles that show cross stratification reaching 25º of dip magnitude. Measured depth is in feet.

Figure 5 - 7: DAVP of the sand package seen in Figure 5-6. Sections of opposed accretions (below 7,685 ft) may indicate a general river flow in an ENE direction.

145

The second example is the sand package that extends from 7,822 ft (2,384 m) to 7,836 ft (2,388 m) (Figure 5-8).The GR shows a sharp basal contact and a fining upward signature indicative of fluvial channel sand/point bars. This is another sand signature found in the same upper (coaly) interval of the Lower Mesaverde. The FMI image shows tool-pull at that depth. Therefore, no bedding planes could be traced nor DAVP plotted. The Lower Mesaverde, also, contains thin sandstone packages with GR profiles which can be described to be coarsening upward at the base of the sand and then fining upward on top. Three sandstones packages (Figure 5-9) found between 8,098 ft (2,468 m) to 8,120 ft (2,475 m) reflect this type of GR signature. Note the chaotic dip pattern and the range of the dip magnitude (3º to 35º). Such sand packages were the toughest to interpret as they show a combination of various features. Interpreted by Longman and Koepsell (2005) to belong to the Middle Neslen fluvial interval of the Lower Mesaverde, these were also deposited in an environment still affected by a coastal-plain depositional environment. The uppermost sand package was interpreted to be a crevasse splay, whereas the lower two can be sections of point bars as the GR at the base is less gradational compared to the GR of the uppermost sand. Also, their fining upward GR is more prominent. The DAVP of the same interval (Figure 5-10) does not reflect much on the flow direction as it is difficult to differentiate the lateral (opposed) accretion from the channel-fill accretion. The sand from depth 8,818 ft (2,688 m) to 8,836 ft (2,693 m) (Figure 5-11) was interpreted as a fining upward fluvial deposit. Interpreting the DAVP (Figure 5-12), we may be looking at opposed accretions (SW and NE), indicating that the water flow direction was toward the SE. Two sand packages (Figure 5-13) have a low dip magnitude sand accretion towards the west in the upper sand, and a relatively higher dip angle in the lower sand with accretion towards the NE (Figure 5-14). The former type of accretion can be lateral accretion in a meandering river section with water flowing towards the NNW The lower sand package can be braided-stream deposits as the higher dip magnitude can suggest higher energy.

146

147

Figure 5 - 8: Example of a point bar from well NBU-222 (refer to text). The FMI image was of poor quality in this zone due to tool pull. Measured depth is in feet.

148

Figure 5 - 9: A section from well NBU-222 of a sand package from the Middle Neslen interval of the Lower Mesaverde. Measured depth is in feet.

Figure 5 - 10: Vector plot of the sand layers seen in Figure 5-9. Lateral and channel-fill accretion directions could not be differentiated on the plot.

149

150

Figure 5 - 11: Channel-fill deposits (8,807-8,817 ft) on top of a lower shoreface sand (8,821-8,848 ft) from well NBU-222.

Figure 5 - 12: The DAVP of amalgamated sand found at depth ranging from 8,818 ft to 8,836 ft.

151

152

Figure 5 - 13: Two sand packages from well NBU-222 that show a low dip sand accretion towards the W, in the upper package (7,958-7,971 ft), and NE in the lower package (7,973-7,984 ft).

Figure 5 - 14: The DAVP of both sand layers of Figure 5-13.

153

5.1.5. Facies Proportions Well NBU-222 covers a measured depth interval that ranges from 6,920 ft (2,109 m) to 9,574 ft (2,918 m). The tops of the Wasatch and Dark Canyon Formations and the Upper Mesaverde Group are not covered by image log. Table 5-2 shows the facies proportions in well NBU-222. The fluvial sand encountered was lumped into braided stream and meandering stream deposits (lateral point bars). Those were found to be 3% and 26%, respectively, of the total well interval. Crevasse splays formed 5% of the total logged intervals. Most of the splays were thin sand bodies (generally of low porosity). Shale formed 55% of the total facies encountered in the well and shallow marine sand or shoreface sand constituted 10% only. Coal was negligible, forming only 0.24% of the total facies present. Almost 3% of the total well interval was not identified or classified into a facies.

5.2. Well Two: Bonanza 4-6 Well Bonanza 4-6 (BON 4-6) is located in the Bonanza field in Utah, SE-NE-SW portion of Section 4 (T10S and R23E) (Figure 1-2). The FMI log covered a depth interval ranging from 6,800 ft (2,073 m) to 8,510 ft (2,594 m) of measured depth. A total of 3,382 bedding planes were picked in well BON 4-6. 2,441 bedding planes were picked in sand beds and grouped in the Sedimentary-Bedding dip set, 111 were interpreted as erosional/scour surfaces, and 830 bedding planes were picked in shale and grouped in the Bed-Boundary dip set (Figure 5-15).

5.2.1. Structural Dip Removal The Bed-Boundary dip set has a mean dip magnitude of 2.24 degrees and mean dip azimuth of 300 degrees (Figure 5-15). To restore the shale beds back to their original horizontal configuration, a structural dip of 2.9 degrees and 317 degrees of dip azimuth (as obtained from plotting the shale beds on a Schmidt plot) has been removed from all

154

155

Table 5 - 2: Facies proportions in well NBU-222 as percentages of the total zone thicknesses.

Figure 5 - 15: A stereonet plot, rose diagram, and dip histogram of the various bedding plane dip sets in well BON 4-6 before structure removal. The “Bed_Boundary” dip set contains bedding planes picked in shaly (high GR) intervals, the “Sedimentary” dip set contains bedding planes picked in sand layers (low GR), and the “Erosional_scour” dip set contains bedding planes interpreted to be erosional surfaces. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

156

features picked on the FMI image. New dip sets were recomputed and presented in terms of relative dip. To create new ASCII files of the recomputed dip sets, the “Dip Removal” module of Geoframe was used. The new sets created are: 1) the “Sedimentary-BeddingSDR dip” set, and 2) the “Bed-Boundary-SDR dip” set. The new sedimentary dip set was used to construct the DAVPs of various sand-packages.

5.2.2. Cumulative Dip Plot A CDP (Figure 5-16) was created for the bedding planes picked in shaly sections in well BON 4-6. Few minor inflection points were seen on the CDP (Figure 5-16); these correspond to the following measured depths: 6,975 ft (2,126 m), 7,063 ft (2,153 m), 7,696 ft (2,346 m), 7,980 ft (2,432 m), 8,000 ft (2,438 m), and 8,166 ft (2,489 m). The FMI image is not clear at 6,975 ft (2,126 m) and 7,063 ft (2,153m) due to tool pull and tool rotation, respectively. At 7,696 ft (2,346 m), a chaotic dip pattern is seen in shale with the dip magnitude ranging between 2 and 28 degrees. Both 7,980 ft (2,432 m) and 8,000 ft (2,438 m) are depths that correspond to surfaces separating horizons of different dip azimuths. Both can indicate micro-fault planes. A major inflection point is seen at sampling point 517, which corresponds to a measured depth of 8,283 ft (2,525 m). It is the top of a fining upward sand package overlain by a 4-foot shale layer. An erosional surface was traced on the FMI image at 8,285 ft (2,525 m) separating sedimentary beds with different dip azimuths. The shale beds in well BON 4-6 were plotted in the structural modules of Geoframe, StrucView. An enhanced cross-section (Figure 5-17) trending 260º from north was produced using the similar-fold model (proprietary of Schlumberger). The cross section shows potential fault planes or fold axes at the following depths: 7,926 ft (2,416 m), 7,210.5 ft (2,198 m), 7,550 ft (2,301 m), 7,976 ft (2,431 m), and 8,163 ft (2,488 m). Two of these depths, 7,976 ft (2,431 m) and, more significantly, 8,163 ft (2,488 m), match closely with inflection points seen on the CDP (Figure 5-16).

157

Figure 5 - 16: Cumulative dip plot of the bedding planes in well BON 4-6 after a 90 GAPI cutoff. The arrows point to inflection points and indicate their depths in feet. Data sets in pink, yellow, cyan, and purple represent bedding planes with dip azimuth ranging from 0 to 90, 90 to 180, 180 to 270, and 270 to 360 degrees from north, respectively. (MVL = Lower Mesaverde Group).

158

Figure 5 - 17: A cross section showing the shale bedding planes in well BON 4-6. Few planes (traced in black) were interpreted to be potential micro-fault planes. (Data plotted using the “StrucView” module of the Geoframe software package; Courtesy of Schlumberger).

159

5.2.3. Dip Azimuth Vector Plot The DAVP (Figure 5-18) shows major inflections points at 7,532 ft (2,296 m) and 7,772 ft (2,369 m). At both inflection points, the slope of the plot changes dramatically. The shallower depth is close to a potential fault plane seen at 7,550 ft (2,301 m) on the cross section (Figure 5-17). Minor inflection points are seen between 7,772 ft (2,369 m) and 8,253 ft (2,515 m) (purple and cyan segments of the DAVP in Figure 5-18).

5.2.4. Facies Examples from Well Bonanza 4-6 Few facies examples from well BON 4-6 will be presented before summarizing the facies proportions of the whole well. The remaining sand packages, the FMI image log and the DAVPs will be provided on the attached data CD-ROM. A table with the interpretation of every sand package in the well is provided in Appendix A. An example of a fining-upward channel fill is seen at a depth that extends from 7,570 ft (2,307 m) to 7,580 ft (2,319 m) (Figure 5-19). Longman and Koepsell (2005) interpreted this sand to be part of the upper coal-bearing zone of the Lower Mesaverde Group. Formation correlation as done by Kerr McGee Corporation (Jerry Cuzella, personal communication) places this same sand package into the Lower Mesaverde as well. The DAVP of the sand package (Figure 5-20) shows three major segments interpreted to be lateral accretion. Going from the bottom of the interval up, the accretions are in a SE, SSW, and NW direction. These indicate river flow direction toward the NE, SE, and NE, respectively. A non-fluvial sand environment is seen at the depth interval that ranges from 8,171 ft (2,491 m) to 8,182 ft (2,494 m) (Figure 5-21), where Longman and Koepsell (2005) noted burrows on the FMI image. The DAVP of the sand (Figure 5-22) shows bedding accretion to the south with planar cross stratifications of a dip magnitude not exceeding 10 degrees. This sand was interpreted to be part of the Lower Neslen Formation deposited as a lagoon washover fan (Longman and Koepsell, 2005). Washover fans form when a barrier island gets washed over by an increase in water level and/or 160

Figure 5 - 18: A dip azimuth vector plot of all shale beds in well BON 4-6. (MVL = Lower Mesaverde Group).

161

162

Figure 5 - 19: A fining-upward meandering channel fill in well BON 4-6. The sand is part of the upper coal-bearing zone of the Lower Mesaverde Group. The “Quicklook” interpretation shows an interlayering of fluvial sand and carbonaceous sand in the interval between 7,574 ft and 7,580 ft.

Figure 5 - 20: The DAVP of the sand in Figure 5-19.

163

164

Figure 5 - 21: A washover sand in well BON 4-6. Depth is in feet.

Figure 5 - 22: The DAVP of the lagoon washover sand of Figure 5-21.

165

large waves often associated with storms. Another example of a washover fan is seen overlying a 1-foot coal bed (Figure 5-23). The DAVP of this fan (Figure 5-24) shows a subtle sand accretion toward the east below 8,252 ft (2,525 m). The sand is interpreted to be part of the Lower Neslen Formation (Longman and Koepsell, 2005) or more generally, part of the Lower Mesaverde Group (Kerr McGee Corporation correlation). The same discussion applies to the sand interval that extends from 8,286 ft (2,526 m) to 8,295 ft (2,528 m) (Figure 5-25). The only difference is that the lower fan sand is the product of multiple accretion events. A lower interval that accretes toward the NE direction, a middle interval accreted toward the SW, and an upper interval accreted toward the NE are seen (Figure 5-26). A marine (lower shoreface) sand package in well BON 4-6 (Figure 5-27) is seen in the Lower Sego Formation from 8,342 ft (2,543 m) to 8,372ft (2,552 m) (Longman and Koepsell, 2005). The sand (Figure 5-28) shows two retrogradational (back-stepping) events toward the west (8,375 ft [2,553 m] to 8,369 ft [2,551 m]) and SW direction (8,369 ft [2,551 m] to 8,348 ft [2,545 m]). Also, a progradation event (8,348 ft [2,545 m] to 8,340 ft [2,542 m]) is seen toward the NW. The shoreline is believed to have been at 90º from the progradation/retrogradation directions mentioned above.

5.2.5. Facies Proportions The image log in well BON 4-6 does not cover the tops of the Wasatch Formation, the Dark Canyon Formation, and the Upper Mesaverde Group (MVU). Table 5-3 shows the facies proportions in the well. The fluvial sand encountered in the well was lumped into braided stream and meandering stream deposits, which were found to be 5% and 32%, respectively, of the total well interval. Crevasse splays formed 5% of the total logged interval. Most of the splays were thin sand bodies (of low porosity in general). Shale formed 46% of the total facies encountered in the well. The well was interpreted to have only 5% and 3% of its interval thickness as shoreface sand (lower and upper) and transition environment sand, respectively. The transition environment included lagoonshoreline deposits (>1%) and washover fans (<2%). Coal was negligible, forming only 0.7% of the total thickness. Less than 3% of the total well interval was not identified or

166

167

Figure 5 - 23: A washover fan/lagoonal shoreline sandstone from well BON 4-6. Depth is in feet.

Figure 5 - 24: DAVP of the washover fan seen in Figure 5-23.

168

169

Figure 5 - 25: A washover fan sand in well BON 4-6. Depth is in feet.

Figure 5 - 26: The DAVP of the washover fan sand seen in Figure 5-25. Note the multiple accretion events.

170

171

Figure 5 - 27: A thick amalgamated shoreface sand section in the Lower Sego Formation in well BON 4-6. Depth is feet.

Figure 5 - 28: Vector plot of a lower shoreface sand in the Lower Sego Formation in well BON 4-6.

172

173

Table 5 - 3: Facies proportions in well BON 4-6 as percentages of the total zone thicknesses.

classified into a facies. Table 5-3 shows the proportion of all the above mentioned facies in each of the various zones in the well.

5.3. Well Three: Pawwinnee 3-181 Well Pawwinnee 3-181 is located in the Pawwinnee field in Utah, NW-NW-NW portion of Section 3, (T9S and R21E) (Figure 1-2). Two logging runs were conducted to cover the interval of interest extending from 9,468 ft (2,886 m) to 12,222 ft (3,725 m) of measured depths with 229 ft (70 m) of overlap between both runs. The logged sections will be referred to as the deeper interval and the shallower interval. A total of 3,774 bedding planes were picked in sand and grouped in the Sedimentary-Bedding dip set, and 1,331 bedding planes were picked in shale and grouped in the Bed-Boundary dip set (Figures 5-29 and 5-30). 100 planes were interpreted to be erosional/scour surfaces.

5.3.1. Structural Dip Removal The Bed-Boundary dip set has a mean dip magnitude and mean dip azimuth of 3 degrees and 316 degrees, respectively, in the shallower log interval, and 3.55 degrees 305 degrees, respectively, in the deeper log interval. Both structural dips (3/316 and 3.55/305) were removed to restore the shale beds back to their original horizontal configuration (Figures 5-29 and 5-30). All features picked on the image were recomputed and presented in terms of relative dip. The new sets created are: 1) the “Sedimentary-Bedding-SDR dip” set, and 2) the “Bed-Boundary-SDR dip” set. The bedding plane dips and azimuths of the new dip sets were used to prepare the DAVPs of the sand packages, some of which are given as examples in the following sections.

174

Figure 5 - 29: A stereonet plot, rose diagram, and dip histogram of the various bedding plane sets of the shallower interval of well Pawwinnee 3-181. The “Bed_Boundary” dip set contains bedding planes picked in shaly (high GR) intervals, the “Sedimentary” dip set contains bedding planes picked in sand layers (low GR), and the “Erosional_scour” dip set contains bedding planes interpreted to be erosional surfaces. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

175

Figure 5 - 30: A stereonet plot, rose diagram, and dip histogram of the various bedding plane sets in the deeper interval of well Pawwinnee 3-181. The “Bed_Boundary” dip set contains bedding planes picked in shaly (high GR) intervals, the “Sedimentary” dip set contains bedding planes picked in sand layers (low GR), and the “Erosional_scour” dip set contains bedding planes interpreted to be erosional surfaces. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger).

176

5.3.2. Cumulative Dip Plot A CDP is created for the bedding planes picked in shale sections in well Pawwinnee 3-181. Gamma ray (GR) values were interpolated to the same depths where bedding planes were picked, and a GR cutoff of 80 GAPI (Figure 5-31) was applied to the data to filter out bedding planes picked in clean sandy layers. The CDP (Figure 5-31) shows minor inflection points at 9,952 ft (3,033 m), 10,002 ft (3,049 m), 10,150 ft (3,094 m), 10,355 ft (3,156 m), 11,196 ft (3,413 m), and 12,078 ft (3,681 m). A major inflection point is seen at 11,083 ft (3,378 m) (Figure 5-31). Plotting all shale beds in the well in StrucView, enhanced cross sections (Figures 5-32 and 5-33) trending 145º and 315º from north were drawn using the similar-fold model (proprietary of Schlumberger). The StrucView cross sections did not show any micro-fault/fault planes.

5.3.3. Dip Azimuth Vector Plot The DAVP of all the bedding planes picked in shale sections in well Pawwinnee 3-181 is shown in Figure 5-34. The DAVP was constructed to show inflection points at depths where unconformities, slumps, or faults may be present. The plot of the shallower interval of the well (Figure 5-34A) is too chaotic to be fully trusted. Few inflection points are seen at roughly 9,735 ft (2,967 m), 9,926 ft (3,026 m), and 10,497 ft (3,200 m). Only the second point mentioned is close to a minor inflection point picked at 9,952 ft (3,033 m) on the CDP (Figure 5-31). A minor unconformity may be found around that same depth. The DAVP of the deeper interval (Figure 5-34B) of the well shows only one significant inflection point between 12,000 ft (3,658 m) and 12,221 ft (3,725 m) (yellow segment in Figure 5-34B) that may be the same inflection point seen at 12,078 ft (3,681 m) on the CDP (Figure 5-31).

177

Figure 5 - 31: Cumulative dip plot of the bedding planes in well Pawwinnee 3-181 after an 80 GAPI GR cutoff. The arrows point to the inflection points and indicate their depths in feet. Data sets in pink, yellow, cyan, and purple represent bedding planes with dip azimuth ranging from 0 to 90, 90 to 180, 180 to 270, and 270 to 360 degrees from north, respectively. (MVL = Lower Mesaverde Group, CGATE = Castlegate Formation, and MN = Mancos Shale).

178

Figure 5 - 32: A cross section showing the shale bedding planes of the upper interval of well Pawwinnee 3-181. (Data plotted using the “StrucView” module of the Geoframe software package; Courtesy of Schlumberger).

179

Figure 5 - 33: A cross section showing the shale bedding planes of the lower interval of well Pawwinnee 3-181. (Data plotted using the “StrucView” module of the Geoframe software package; Courtesy of Schlumberger).

180

(A)

(B)

Figure 5 - 34: The DAVP of the bedding planes picked in shale sections in the (A) shallower interval and (B) deeper interval of well Pawwinnee 3-181.

181

5.3.4. Facies Examples from Well Pawwinnee 3-181 A few facies examples from well Pawwinnee 3-181 will be presented before summarizing the facies proportions of the whole well. The remaining sand package, the FMI image log and the DAVPs will be provided on the attached CD-ROM. A table with the interpretation of every sand package in the well is provided in Appendix A. An example of a fluvial blocky sandstone is seen from 9,590 ft (2,923 m) to 9,606 ft (2,928 m) (Figure 5-35). Longman and Koepsell (2005) and the correlation done by Kerr McGee Corporation (Jerry Cuzella personal communication, 2006) placed this interval in Upper Mesaverde Group. The DAVP of the same sand (Figure 5-36) shows opposed accretion in NE and NW directions. This may indicate amalgamated lateral point bar deposits in meandering streams with a general flow direction towards the NE and SE. Another fluvial amalgamated sand is seen in Figure 5-37. The DAVP (Figure 538) shows some minor opposed accretions that may reflect some trough cross stratifications in a lateral point bar. The dip magnitude exceeds 20 degrees in some sections. Accretion towards the east can be river bed accretions. In addition to the accretion, the DAVP shows some chaotic dip patterns indicative of braided stream deposits. A sand package that extends from 10,647 ft (3,245 m) to 10,670 ft (3,252 m) (Figure 5-39) is interpreted by Longman and Koepsell (2005) to be both washover fan sand deposits and part of the Lower Neslen Formation. The DAVP of the washover fan (Figure 5-40) shows a main dip direction towards the SW. The sand that extends from 10,810 ft (3,295 m) to 10,840 ft (3,304 m) (Figure 541) is a lower to upper shoreface sand from the Lower Sego Formation. Longman and Koepsell (2005) interpreted some features on the FMI image to be burrows and shale clasts. The DAVP (Figure 5-42) shows a series of different events that may be retrogradation towards the SW and progradation towards the E and SE. Similarly, the interval that extends from 10,854 ft (3,308 m) to 10,875 ft (3,315 m) (Figure 5-43) is another upper shoreface sand from the Upper Castlegate Formation. The DAVP of the sand (Figure 5-44) shows two main progradational accretions in NE and E directions.

182

183

Figure 5 - 35: Fluvial deposits in the Upper Mesaverde in well Pawwinnee 3-181. Depth is in feet.

Figure 5 - 36: DAVP of the fluvial sand in Figure 5-35 showing opposed accretion events indicating a river that flowed in a NE and SE directions.

184

185

Figure 5 - 37: Amalgamated fluvial channel deposits in well Pawwinnee 3-181. Depth is in feet.

Figure 5 - 38: DAVP of the sand section seen in Figure 5-37.

186

187

Figure 5 - 39: A sand package from well Pawwinnee 3-181. The sand is interpreted as a washover fan from the Lower NeslenFormation by Longman and Koepsell (2005). Depth is in feet.

Figure 5 - 40: DAVP of the washover fan sand of Figure 5-39. Note the subtle accretion direction toward the SW.

188

189

Figure 5 - 41: A lower to upper shoreface sand seen in the Lower Sego Formation in well Pawwinnee 3-181. Depth is in feet.

Figure 5 - 42: DAVP of the shoreface sand seen in Figure 5-41. The shoreline is expected to have been perpendicular to the accretion directions (refer to text).

190

191

Figure 5 - 43: An upper shoreface sand from the Upper Castlegate Formation in well Pawwinnee 3-181. Depth is in feet.

Figure 5 - 44: DAVP shows episodes of progradation in NE, E and ENE directions, and retrogradation to the west.

192

Well Pawwinnee 3-181 extended deep enough to penetrate the Mancos Shale. Sand sections in the Mancos Shale (Figure 5-45 and 5-46) extending from 11,800 ft (3,597 m) to 11,822 ft (3,603 m) and 12,044 ft (3,671 m) and 12,056 ft (3,675 m) (Figure 5-47) were interpreted to be a silty sand package and turbidite deposits, respectively, by Longman and Koepsell (2005).

5.3.5. Facies Proportions The image log in well Pawwinnee 3-181 does not cover the tops of the Wasatch Formation, the Dark Canyon Formations, or the Upper Mesaverde Group (MVU). Table 5-4 summarizes the facies proportions in the well. The fluvial sand encountered in the well was lumped into braided stream and meandering stream deposits (lateral point bars). Those were found to be 1% and 16%, respectively, of the total well interval. Crevasse splays formed 4% of the total logged intervals. Most of those were thin sand bodies of poor quality (low porosity and shaly) sand. Shale formed 58% of the total facies encountered in the well. Shoreface sand accounted for 14%. In addition to the previously mentioned facies, well Pawwinnee 3-181 has facies not encountered in any of the other three wells. These are marine silty sands and turbidite deposits from the Mancos Shale. Washover fan, marine silty sand, and turbidites accounted for, approximately, 1% each of the total log interval. Coal was negligible, forming only 0.6% of the total facies present. 3% of the sand were not identified or classified into a facies. The undefined sandstones were mainly from the Mancos Shale and the Lower Mesaverde Formation.

5.4. Well Four: Kennedy Wash Federal Unit 16-1 Well Kennedy Federal Unit (KWFU) 16-1 is located in the KWFU field in Utah, in the NW-NW portion of Section 16 (T9S and R21E) (Figure 1-2). The FMI log covered an interval that ranges from 8,490 ft (2,588 m) to 10,991 ft (3,350 m) of measured depth.

193

194

Figure 5 - 45: Planar cross stratification in a silty sand package from the Mancos Shale in well Pawwinnee 3-181. Depth is in feet.

Figure 5 - 46: DAVP of the silty sand in the Mancos Shale seen in Figure 5-45.

195

196

Figure 5 - 47: Turbidite deposits from the Mancos Shale in well Pawwinnee 3-181. Depth is in feet.

197

Table 5 - 4: Facies proportions in well Pawwinnee 3-181 as percentages of the total logged zone thicknesses.

A total of 2,687 bedding planes were picked in sand and grouped in the Sedimentary-Bedding dip set, and 724 bedding planes were picked in shale and grouped in the Bed-Boundary dip set (Figures 5-48). 30 planes were interpreted to be erosional/scour surfaces.

5.4.1. Structural Dip Removal The Bed-Boundary dip set has a mean dip magnitude of 3.27 degrees and mean dip azimuth of 318 degrees (Figure 5-48). This structural dip was removed to restore the shale beds back to their original horizontal configuration. All features picked on the image were recomputed and presented as relative dip. The new sets created are: 1) “Sedimentary-Bedding-SDR Dip” set, and 2) “Bed-Boundary-SDR Dip” set. They were used to prepare the DAVPs for various sand packages in the well. The bedding plane dips and azimuths of the dip sets before structural removal are used to prepare the CDP and the DAVP of the shale beds.

5.4.2. Cumulative Dip Plots A CDP (Figure 5-49), was created for the bedding planes picked in shale (before structure removal) sections in well KWFU 16-1. Gamma ray (GR) values were interpolated to the same depths where bedding planes were picked, and a GR cutoff of 80 GAPI was applied to the data to filter out bedding planes picked in clean sandy layers. In the well, no major or significant inflection point was seen. Still, many minor ones are seen at the following depths: 9,150 ft (2,789 m), 9,164 ft (2,793 m), 9,538 ft (2,907 m), 9,621 ft (2,933 m), 9,752 ft (2,972 m), 9,811 ft (2,990 m), 9,962 ft (3,036 m), 10,091 ft (3,076 m), 10,153 ft (3,095 m), 10,283 ft (3,134 m), 10,301 ft (3,140 m), 10,384 ft (3,165 m), 10,401 ft (3,170 m), 10,470 ft (3,191 m), 10,510 ft (3,204 m), 10,580 ft (3,225 m), and 10,815 ft (3,296 m). Plotting all shale beds in well KWFU 16-1 in “StrucView,” an enhanced cross section (Figure 5-50) that trends 135º from north was drawn using the similar-fold model

198

Figure 5 - 48: A stereonet plot, rose diagram, and dip histogram of the various bedding plane sets in well KWFU 16-1. (Data plotted using the “BorView” module of the Geoframe software package; Courtesy of Schlumberger). The “Bed_Boundary” dip set contains bedding planes picked in shaly (high GR) intervals, the “Sedimentary” dip set contains bedding planes picked in sand layers (low GR), and the “Erosional_scour” dip set contains erosional surfaces.

199

Figure 5 - 49: Cumulative dip plot of the bedding planes in well KWFU 16-1 after an 80 GAPI cutoff. The arrows point to the inflection points and indicate their depths in feet. Data sets in pink, yellow, cyan, and purple represent bedding planes with dip azimuth ranging from 0 to 90, 90 to 180, 180 to 270, and 270 to 360 degrees from north, respectively. MVL = Lower Mesaverde Group and CGATE = the Castlegate Sandstone.

200

Figure 5 - 50: A cross section showing the shale bedding planes in well KWFU 16-1. Only one plane is seen to be a potential fault plane at 10,231 ft. (Data plotted using the “StrucView” module of the Geoframe software package; Courtesy of Schlumberger).

201

as the monocline model did not show enough continuity nor revealed structural features. StrucView shows one potential fault plane at 10,231 ft (3,118 m) (Figure 5-50). This depth does not closely match any inflection point seen on the CDP. However, plotting the DAVP of the sand interval extending from 10,238 ft (3,121 m) to 10,338 ft (3,151 m) (Figure 5-51), a major inflection point was seen at 10,231 ft (3,149 m). The inflection point indicated a shift in the structural dip direction from a SW to E direction.

5.4.3. Dip Azimuth Vector Plot The DAVP (Figure 5-52) was plotted for all of the bedding planes picked in shale sections. It is another plot that may show inflection points at depths where unconformities, slumps, or faults are present. The DAVP of KWFU 16-1, like the well’s CDP, shows no significant inflection point at any depth.

5.4.4. Facies Examples from Well KWFU 16-1 A few examples of each environment will be presented. The FMI image logs, the remaining sand packages, and their DAVPs will be provided on the attached CD-ROM. A table with the interpretation of every sand package in the well is provided in Appendix A. A fining-upward channel fill extends from 8,588 ft (2,618 m) to 8,616 ft (2,616 m) (Figure 5-53) and is interpreted to be part of the Upper Mesaverde Group (Longman and Koepsell, 2005 and Kerr McGee Corporation). The DAVP of this sandstone (Figure 5-54) shows opposed accretion toward the north and south, suggesting a river flowing to the east. Another example of a fluvial channel sand extends from 9,000 ft (2,743 m) to 9,026 ft (2,752 m) (Figure 5-55). The opposed accretion seen in the DAVP of the sand (Figure 5-56) shows three major segments indicating (lateral) accretion events. Going from the base up, the accretion directions are toward the WSW, NE, and W (with a minor topmost accretion toward the east). These indicate river segments flowing towards the

202

Figure 5 - 51: A dip-azimuth vector plot of sand interval extending from 10,226 ft to 10,238 ft showing an inflection point at 10,231ft.

203

Figure 5 - 52: The dip azimuth vector plot of all the bedding planes picked in shale layers in well KWFU 16-1 before structural dip removal. The plot does not show any significant inflection point. (MVL = Lower Mesaverde Group and CGATE = the Castlegate Sandstone).

204

205

Figure 5 - 53: A fining-upward sand package from the Upper Mesaverde Group in well KWFU 16-1 interpreted as fluvial channel deposits. Depth is in feet.

Figure 5 - 54: DAVP of the fluvial sand in Figure 5-53 showing mainly accretions toward north and south.

206

207

Figure 5 - 55: A fining-upward fluvial sand from the Upper Mesaverde Group in well KWFU 16-1. Depth is feet.

Figure 5 - 56: DAVP of the fluvial sand in Figure 5-55.

208

SE, E, and N, respectively. The sand of some units in the Upper Mesaverde Group are interpreted to be braided stream deposits (Longman and Koepsell, 2005) such as the interval that extends from 9,344 ft (2,848 m) to 9,380 ft (2,859 m) (Figure 5-57). This interval shows a sharp basal GR contact and fining-upward sand. The DAVP of the same sand (Figure 5-58) shows the typical chaotic plot pattern expected in braided stream channel fills although the GR does not show the expected blocky signature of braided stream deposits. A Castlegate shoreface sand extends from 10,740 ft to 10,678 ft (Figure 5-59). Note the gradational GR basal contact. The DAVP of the same sand (Figure 5-60) shows retrogradation or back-stepping of deposits (knowing that the seaway was to the east) in a NW direction and progradation events in NE and N directions.

5.4.5. Facies Proportions Table 5-5 summarizes the facies proportions of well KWFU 16-1. The fluvial sand encountered in the well was lumped into braided stream and meandering stream deposits. Those were found to be 5% and 26%, respectively, of the total well interval. Crevasse splays formed >2% of the total logged intervals. Most of the splays were thin sand bodies of poor quality (low porosity and shaly) sand. Shale formed 47% of the total facies encountered in the well. The shale, except the sections between the shoreface sand, is considered to be (continental) floodplain shale. Coal was negligible in well KWFU 161. Washover fans and coastal plain deposits were 0.8% and 2%, respectively. Shoreface sand accounted for 7% of the total logged interval, 10% of which was not identified or classified into facies because of either poor FMI image and/or bad hole conditions.

5.5. Discussion The facies proportions as obtained from the entire logged intervals in all 4 wells are summarized in Figure 5-61. 55% of the logged intervals in all wells is made of both continental and marine sand. Henshaw (2005) reported comparable facies proportions

209

210

Figure 5 - 57: Braided stream sand deposits from the Upper Mesaverde Group in well KWFU 16-1. Depth is in feet.

Figure 5 - 58: DAVP of the braided stream channel fill of Figure 5-57. The chaotic accretion pattern is typical of braided stream deposits.

211

212

Figure 5 - 59: A shoreface sand section from the Castlegate Formation in well KWFU 16-1. Depth is in feet.

Figure 5 - 60: DAVP of the shoreface sand in Figure 5-59. Note the retrogradation (W, NW and N) and the progradation (NNE).

213

214

Table 5 - 5: Facies proportions in well KWFU 16-1 as percentages of the total logged zones.

from the Riverbend area in the Uinta basin. His outcrop data contained sections from the “fluvial Wasatch and Mesaverde deposits, braided stream deposits of the Castlegate and coastal plain and marine deposits of the Blackhawk Formations” (Henshaw, 2005). His proportions of channel sand, crevasse splays, floodplains, coastal plain (coal), and marine shoreface were >30%, <10%, >50%, <1%, and 5%, respectively. When various zones are looked at separately (Figure 5-62), the channel sand decreases deeper in the wells. The crevasse splays constitute around 5% of both the Upper Mesaverde (MVU) and the Lower Mesaverde (MVL) sections. It becomes less than 1% in the Castlegate Formation (CGATE). However, the shoreface sand, not seen in the MVU, increases in proportion as we go from MVL to CGATE. When we compare the environments of deposition or their combination in each of the zones (MVU, MVL, and CGATE), we notice a dominant fluvial environment in the MVU and a dominant marine environment in the CGATE. MVL, on the other hand, shows a transition of environments where lagoon deposits and washover fans are seen. The facies vertical distribution and proportions obtained in well NBU-222 in both MVU and MVL will be used as input into Petrel to create a facies log and a facies model around the well (Chapter 6).

215

Figure 5 - 61: Facies proportions in MVU, MVL, and CGATE from all 4 wells.

216

(B) (A)

(C)

Figure 5 - 62: Facies proportions calculated from all 4 wells in (A) MVU, (B) MVL, and (C) the CGATE Formation.

217

CHAPTER SIX

3D FACIES MODEL

After computing the vertical facies proportions in well NBU-222 (Chapter-5), a 3dimensional (3D) geologic model was built using Petrel to show the sand distribution around the well. Well NBU-222 was chosen for modeling because it has the image log with the least gas entry and tool pull. The well had a good coverage of the Upper and Lower Mesaverde Group sections and the Castlegate Formation. Also, we could find many wells around NBU-222 with known formation tops. This helped control the shape of the horizons created to separate the various modeled zones (later in the chapter). The main objectives of the Petrel model are to understand the subsurface sand distribution, to calculate the total number of sand bodies intersected in each of the infilldrilling scenarios and, when possible, the number of sand bodies intersected by more than 1 well, and to recommend the best drilling scenario to adopt in a similar environment. This chapter discusses how the Petrel model was built using an object-based (stochastic) modeling approach. Also, it shows how the sand dimensions and orientations were chosen. An area containing the 4 interpreted wells (large polygon in Figure 6-1) could not be modeled, as it would contain tens of millions of cells. A much smaller area (discussed later) around well NBU-222 was modeled instead. Pseudo-wells distributed around well NBU-222 were used to check the validity of the model by checking the N/G ratio and the average number of sand bodies intersected by any well around the subject well NBU-222. An application to analyze connectivity of sand cells with well trajectories is provided in Appendix B. The application determined the total number of “geobodies” (defined in Appendix B) and sand volumes intersected in each infill-drilling scenario. The efficiency of various infill-drilling scenarios intersecting and draining fluvial sand bodies is demonstrated at the end of the chapter.

218

Figure 6 - 1: A map view to show the locations of the four wells (labeled) interpreted in the study, and the wells in Table 6-2 that were used to create the various surfaces seen in Figure 6-3. Green arrow points north.

219

6.1. Model Grid Design Input (ASCII) files were prepared to enter the well specifications (well locations, well-head elevations, well-bottom depths, and formation tops) in Petrel (Figure 6-1; Table 5-1). A synthetic facies log (Figure 6-2) is produced in Petrel to depict the image log interpretation of well NBU-222. In order not to lose the vertical resolution when the facies log is upscaled, the grid cell thickness is chosen to be 3 ft (0.91 m). Formations tops were used as control points to create horizons that connect the same stratigraphic events (formation tops) in the studied wells. Having only 4 scattered wells provided insufficient control to create the horizons. Twenty-two wells (Table 6-1) were added around well NBU-222 (smaller polygon in Figure 6-1) for more control on the shape of the needed horizons. The formation tops in the added wells are not published information and will not be provided. Some of the created surfaces/horizons can be seen in Figure 6-3. They divide the Mesaverde into various zones, which were modeled separately. Another reason for modeling a small area around well NBU-222 was the uncertainty whether facies logs from scattered wells can provide enough control on a model. Petrel, as will be shown, honors any input, no matter how scarce the control/input points are in the modeled area. Therefore, a 1.16 mi (1.87 km) by 1.17 mi (1.18 km) area is modeled. This area is enclosed by the polygon in Figure 6-4 and has a surface area of 1.36 mi2 (3.5 km2). Within this surface area, it was possible to have the grid cells of 3 ft (0.91 m) in thickness, thereby preserving the thin sand layers interpreted as crevasse splays in the facies log. After creating the horizons around well NBU-222 (Figure 6-3), a function in Petrel was used to erase the portions of the horizons outside of the modeled polygon and leave only the portions inside the modeled area (Figure 6-4). Those horizons helped shape and control the model grid skeleton (Figure 6-5A). With the horizons created to separate the various modeled intervals and shape the grid skeleton, the 3D grid was created. The cell dimensions are 100 ft (30.48 m) x 100 ft (30.48 m) x 3 ft (0.91 m). Consequently, the model contains a total of 2.64 million (62 x

220

Figure 6 - 2: Well NBU-222 borehole image log transformed into a facies log (track 1) and entered in Petrel. Density, GR, resistivity, and neutron porosity curves are seen in tracks 2, 3, 4, and 5, respectively. (yellow = channel sand, pink = crevasse splays, green = shoreface sand, orange = marine shale, brown = flood plain, black = coal, DKCYN = Dark Canyon Formation, CGATE = Castlegate Formation, and MN1 = Mancos Shale).

221

Figure 6 - 3: The surfaces created that connect the formation tops among the wells and separating the Upper and Lower Mesaverde intervals from the Castlegate Sandstone and the Mancos Shale. The spheres show various formation tops. The Upper Mesaverde was further divided into three intervals. Green arrow points north. (The scale of the figure is not specified as the figure was rotated to show the various surfaces. Depth scale bar covers 6,000 ft).

222

Figure 6 - 4: A polygon that designates the area modeled around well NBU-222. The surfaces inside the polygon helped form and control the model grid skeleton. Green arrow points north. The modeled area shown covers 1.36 mi2 (3.5 km2).

223

(A)

(B)

Figure 6 - 5: (A) The horizons relating formation tops and separating the modeled sections are seen in addition to grid top, mid, and lower skeletons; (B) The various modeled zones. Green arrow points north. Skeleton surfaces seen cover 1.36 mi2 (3.5 km2), and the modeled zones cover 2,091 ft (637 m).

224

Table 6 - 1: The coordinates, Kelly-bushing (KB) elevations, and TD of the four studied wells (bold and underlined) and the 22 wells added around well NBU-222.

61 x 697) cells. Note that 697 cells, stacked vertically, give a total thickness of 2,091 ft (637 m). However, the Mesaverde Group thickness in well NBU-222 is 1700 ft (518 m) only. Petrel tried to interpolate the model to the top of the Mesaverde Group, interpreted to be at 6,531 ft (1,991 m) of measured depth and 389 ft (118 m) above the actual top of the FMI log. This is also the reason why the top skeleton of the grid (Figure 6-5A) is above the MVU-a surface (explained below). The lower skeleton of the grid (Figure 65A) is below the Castlegate Formation top/surface. For the elements proportions and N/G computations, only the Mesaverde Group interval between the top of the FMI log and the top of the Castlegate Sandstone is considered. The various sandstone formations that form the Mesaverde Group in well NBU222 were not separated. Therefore, the Mesaverde Group is referred to as the Upper and Lower Mesaverde. The Upper Mesaverde is further subdivided into three zones: 1) the upper Upper Mesaverde named MVU-a, the middle Upper Mesaverde named MVU-b,

225

and the lower Upper Mesaverde named MVU-c (Figure 6-5B). The sand-package thickness variations and the fining and coarsening-upward sequences (GR signatures) were the criteria used to divide the Upper Mesaverde into MVU-a, MVU-b, and MVU-c. This allowed working with better sand-dimension distributions instead of one set of generalized mean, minimum, and maximum values in the upper 919 ft (281 m) of the well. A less variable sand-thickness distribution is observed in the Lower Mesaverde (MVL). Therefore, this latter interval is treated and modeled as one zone (Figure 6-5B).

6.2. NBU-222 Facies Proportions The facies proportions obtained interpreting the sand packages in well NBU-222 are summarized in Table 5-2. The proportions in Table 5-2 are computed from the whole FMI-logged interval in the well, including the Castlegate Sandstone and the Mancos Shale. In Petrel, only the interval from 6,920 ft (2,109 m) to 8,620 ft (2,627 m) of measured depth is considered. In those 1,700 ft (518 m), the sand proportions were 38%, 12%, and 4% for channel sands, crevasse splays, and shoreface sands, respectively. The shale formed 46% of the modeled section/interval, and there is less than 1% of coal in the well.

6.3. Petrel Input: Sand Element Dimensions and Shapes Choosing the object-based modeling approach requires assigning shapes and dimensions (Figure 6-6) to the elements modeled after specifying their proportions in every modeled section. The different sand elements were assigned different geometrical shapes. Half an ellipse was used to model point bars or channel-fill sand, whether interpreted to be the result of a single-event deposition or an amalgamation/series of events. A circle was the shape used to depict the crevasse-splay sand. Both shoreface sand and coal were modeled as sheet-like bodies, which is likely at the scale of the modeled area.

226

Figure 6 - 6: An example of the needed input data for every facies or modeled element in Petrel.

227

Table 6 - 2: A summary of the mean, minimum, and maximum values of the sand element orientations and dimensions used for the modeled zones.

228

The fluvial sand-element dimensions (Table 6-2) were computed using relationships obtained by Panjaitan (2006). With the sand thickness (x) known, the equations for the sand-body width (y) are y = 30.2 x0.72 for single channel-fill events and y = 39.2 x0.62 for amalgamated or superimposed channel-fill events (Panjaitan, 2006); where units are in meters. Both relationships were adopted to compute the meanderingriver deposit widths in well NBU-222. The thickness seen in the well was corrected for compaction (10% of a sand package thickness was added to account for compaction). The N/G in well NBU-222 is far greater than the N/G of Panjaitan (2006). In his subsurface (only 1 well) and outcrop data, Panjaitan (2006) had N/G of 0.17:1 and 0.15:1, respectively. He argued that only one well may not represent the N/G of the subsurface data, and the “pseudo-wells in the outcrop do not represent vertical wells, but rather deviated wells” (Panjaitan, 2006). His relationships seem to be more applicable to calculate sand-body widths of the Mesaverde Group than the published relationships computed from a wider range of river systems in various locations. Panjaitan (2006) compared and contrasted the sand-body widths computed with the above relations to other published studies like Lorenz et al. (1985) and Bridge and Mackey (1993). He showed that the widths of the outcropping fluvial sand bodies he measured in the Piceance basin by the aerial Light Detection and Ranging (LiDAR) techniques are at least one order of magnitude less than the widths computed from other published relationships. Therefore, for meander stream deposits (amalgamated and single-event deposits), Panjaitan’s (2006) relationships were used. In well NBU-222, the thickness of the sand packages ranges from 3 to 55 ft (0.91 to 16.8 m), with a corresponding width ranging from 129 to 563 ft (39.3 to 171.6 m). An elliptical form (Figure 6-7A) was adopted to depict the map-view shape of the sand bodies rather than the half-circle shape used by Panjaitan (2006) (Figure 6-7B). The reason for using another shape to model the point bars is to reduce the sand-body area in this study and compare the results with Panjaitan’s (2006). Figure 6-7C shows that the sand-body areas in the Petrel model are half of what Panjaitan (2006) would have calculated. The sand-body areas ranged from 0.3 to 5.7 ac (0.0012-0.023 km2). One reason for such an approach is to compare the percentages of sand bodies intersected (in our various infill-drilling scenarios) with the percentages calculated by Panjaitan (2006).

229

Figure 6 - 7: (A) The elliptical shape used for the sand bodies in the Petrel model compared to (B) the circular shape used by Panjaitan (2006) reducing (C) the area of a sand body by half.

230

Panjaitan (2006), as in this study, modeled crevasse splays as circles and provided no relationship to relate the thickness of a crevasse splay to its width or diameter. He only reported ranges (means, minimums and maximums) of thicknesses and widths for the crevasse splay deposits. The thickness of the crevasse splays are used as seen on our FMI log ranging from 0.5 to 6.5 ft (0.15 to 1.98 m). Panjaitan’s (2006) width distribution is used in our facies model because we had no other means to determine the crevasse splay diameters in the Mesaverde Group. The width of crevasse splay reported by Panjaitan (2006) ranges from 45.9 to 1661 ft (14 to 506.3 m), and, therefore, the area ranges from 0.04 to 49.7 ac (0.00015-0.2 km2). The dimensions (mean, minimum, and maximum values) of the braided-river deposits (Table 6-2) are adopted from German (2006), who studied the Upper Williams Fork Formation at Plateau Creek Canyon in the Piceance basin. The thickness of the sand packages ranges from 9 to 54 ft (2.7 to 16.5 m), and their width ranges from 204 to 2565 ft (62.2 to 781.8 m). An elliptical form (Figure 6-7A) was adopted to depict the map-view shape of the braided-stream deposits rather than the half-circle shape used by Panjaitan (2006) (Figure 6-7B). Therefore, the area of the braided-stream deposits ranges from 0.75 to 118.6 ac (0.003-0.48 km2). As mentioned before, shoreface sand and coal were modeled as sheets. A coal sheet was assumed to cover the whole modeled area (Table 6-2). A shoreface-sand sheet was assumed to have a mean width (or more correctly a mean side) of 5,000 feet (1,524 m) (Table 6-2). After deciding on the element dimensions and shapes, knowing that the (flood plain) shale occurs as the background facies in the model, the orientations of the various shapes were needed. With a general knowledge that the rivers during the Cretaceous in Utah flowed from west to east, Lawton (1986) and German’s (2006) studies were referred to. Both authors described generalized paleocurrent directions in meandering and braided sections of the Mesaverde Group, respectively. Lawton (1986) studied fluvial outcrops in various areas in Utah. Our MVU-a, MVU-b, and MUV-c were compared to his upper coaly member of the Mesaverde and the Tuscher Sandstone, the Tuscher Sandstone and upper Price River Formation, and the Farrer Formation, respectively. The mean, minimum and maximum paleocurrent orientations of each zone are summarized in Table

231

6-2. It has been inferred from Lawton’s (1986) paper that the Lower Mesaverde had a range of paleocurrent directions from NNE to SE, and even S. To provide Petrel with paleocurrent directions for MVL, the mean, minimum, and maximum paleocurrent directions used are 95, 65, and 135 degrees from north, respectively. For bodies with no major/long axis (sheets and circles), the paleocurrent direction is specified as a full range from 0 to 360 degrees. Crevasse splays and shoreface sand have paleocurrent mean direction of 137 and 31 degrees from north, respectively. These values were obtained from stereonet plots of the sedimentary picks/features from the FMI image in a corresponding sand section in well NBU-222. The paleocurrent direction of the braidedstream deposits or packages was adopted from German (2006). The mean direction is 92 degrees from north with data ranging from 57 to 130 degrees.

6.4. Petrel Model Results After Petrel was run with all input and specifications, a 3D facies model was produced. The model shows the distribution of the various sand elements with shale specified as the background facies in every modeled section (Figure 6-8). In order to check the validity of the facies model, both the facies proportions and the N/G were checked in 34 vertical pseudo-wells inserted at a 5-ac (0.02 km2) spacing (Figure 6-9) around well NBU-222. First, pseudo-facies logs were computed from the model at various well locations. The pseudo-facies logs are the vertical distribution of the facies elements at a specific well location or along a vertical well path. Second, the N/G ratios were computed in the same wells to obtain average ratios that can be compared to the observed N/G in well NBU-222. Table 6-3 shows the facies proportions from the 34 pseudo-wells around well NBU-222 (Figure 6-9). The computed average of the various element proportions yields a N/G of 54% (± 4.4) compared to 54% in well NBU-222 from all three modeled zones. This result shows that Petrel honors the input data even with only one control/input point at the center of the modeled area. The model showed that out of 26,604 sand bodies of various dimensions in the model, NBU-222 penetrates 81 sand bodies of various thicknesses. The same wells of Table 6-3 were considered to determine the number of sand bodies intersected by each well. Table 6-4 shows that a

232

Figure 6 - 8: The Petrel facies model (yellow = channel sands, pink = crevasse splays, green = shoreface sands, brown = floodplain). Pseudo-wells are inserted in the modeled volume. Green arrow points north. The model seen covers a 1.36 mi2 (3.5 km2) area and 2,091 ft (637 m) of depth.

233

Figure 6 - 9: A map view showing the 34 wells used to compute the average facies proportions and number of sand bodies in Tables 6-4 and 6-5. The wells are distributed at 5-ac (0.02 km2) spacing around well NBU-222.

234

Table 6 - 3: Elements/facies proportions of the Mesaverde Group (Upper and Lower) in 5-ac (0.02 km2) spaced wells around well NBU-222 (Figure 6-9) with the average proportions computed. Well W5-18 is well NBU-222. It is not included in the table or in the computed average proportions.

45.86 44.16 48.52 48.43 38.75 46.90 46.64 49.12 43.33 40.34 45.99 47.19 42.24 40.02 39.98 45.21 52.24 47.41 52.08 42.30 51.59 50.87 47.05 50.69 54.63 40.61 43.97 36.61 40.85 46.15 40.15 45.85 49.63 43.66

Channel Sand 45.12 48.01 42.61 43.56 49.67 43.78 42.37 37.50 46.32 50.13 45.89 42.90 48.33 49.49 47.87 45.39 40.29 43.47 37.95 48.28 32.44 35.78 39.44 36.14 31.16 47.35 41.67 48.93 42.08 39.88 48.07 40.87 35.22 42.86

Crevasse Splay 6.91 5.00 6.39 5.37 7.90 5.10 5.67 7.75 6.06 5.65 4.41 7.63 5.61 6.09 7.04 4.82 3.82 6.07 6.50 6.08 9.72 8.49 8.73 5.76 6.40 5.33 6.81 7.37 8.12 7.60 5.68 6.14 9.61 6.69

Shoreface Sand 1.76 2.47 2.12 2.47 3.51 4.05 5.15 5.08 4.11 3.71 3.53 2.10 3.64 3.87 4.94 4.57 3.32 2.88 3.47 3.15 5.88 4.48 4.59 7.23 7.46 6.71 7.20 6.74 8.95 6.19 5.92 6.96 5.35 6.40

45.56

42.97

6.54

4.42

5.10

1.40

Well

Shale

W5-1 W5-2 W5-3 W5-4 W5-5 W5-6 W5-7 W5-8 W5-9 W5-10 W5-11 W5-12 W5-13 W5-14 W5-15 W5-16 W5-17 W5-19 W5-20 W5-21 W5-22 W5-23 W5-24 W5-25 W5-26 W5-27 W5-28 W5-29 W5-30 W5-31 W5-32 W5-33 W5-34 W5-35 Average Proportions (%): Standard Deviation

235

Coal

N/G

0.35 0.35 0.35 0.18 0.18 0.18 0.18 0.54 0.18 0.18 0.18 0.18 0.17 0.53 0.18 0.00 0.33 0.18 0.00 0.19 0.38 0.37 0.18 0.18 0.36 0.00 0.35 0.35 0.00 0.18 0.19 0.19 0.19 0.39

53.79 55.48 51.13 51.39 61.08 52.92 53.19 50.34 56.49 59.48 53.83 52.64 57.59 59.45 59.85 54.79 47.43 52.41 47.92 57.51 48.03 48.75 52.76 49.13 45.02 59.39 55.68 63.04 59.15 53.67 59.67 53.97 50.18 55.95

4.70

0.23

54.21

1.81

0.14

4.44

Table 6 - 4: Number of the various element bodies penetrated by 34 wells distributed in a 5-ac (0.02 km2) grid spacing around well NBU-222 (Figure 6-9). Well W5-18 is well NBU-222. It is not included in the table nor in the computed average number of sand bodies intersected.

Coal

Shoreface Sand

Total Sand Bodies Intersected per well

34 26 34 28 33 24 29 37 31 26 21 33 28 29 33 24 21 25 31 31 45 42 38 27 32 28 32 32 39 29 24 33 41 29

2 2 2 1 1 1 1 3 1 1 1 1 1 3 1 0 2 1 0 1 2 2 1 1 2 0 2 2 0 1 1 1 1 2

3 4 6 5 5 6 6 9 7 5 5 5 8 6 6 6 2 6 7 5 12 7 12 6 9 8 12 10 10 7 8 13 10 11

74 70 83 71 75 72 83 85 83 70 69 79 78 79 79 66 60 73 82 77 96 88 95 66 82 71 84 87 78 77 73 84 92 82

40

31

1

7

78

5.08

3.73

5.76

0.76

2.72

8.29

73

52

27

6

2

81

Shale Layers

Channel Sands

Crevasse Splays

W5-1 W5-2 W5-3 W5-4 W5-5 W5-6 W5-7 W5-8 W5-9 W5-10 W5-11 W5-12 W5-13 W5-14 W5-15 W5-16 W5-17 W5-19 W5-20 W5-21 W5-22 W5-23 W5-24 W5-25 W5-26 W5-27 W5-28 W5-29 W5-30 W5-31 W5-32 W5-33 W5-34 W5-35

62 55 66 58 56 59 66 68 62 57 57 61 54 57 61 51 53 57 66 63 65 64 69 50 59 53 56 57 56 55 54 54 64 54

37 40 43 38 37 42 48 39 45 39 43 41 42 44 40 36 37 42 44 41 39 39 45 33 41 35 40 45 29 41 41 38 41 42

Average Numbers of bodies intersected per well:

59

Well

Standard Deviation NBU-222

236

well can intersect as many as 96 sand bodies or as few as 60 sand bodies. It is seen that the average numbers of shale layers, channel sand bodies, and coal layers are underestimated compared to their numbers in well NBU-222. On the other hand, the average number of shoreface sand bodies intersected by the 34 wells is overestimated compared to the number of shoreface sand bodies intersected in well NBU-222. However, the computed average number of sand bodies intersected by any well was found to be 78 sand bodies (within 1 standard deviation [a measure of the spread of the values of a distribution or population. It is defined as the root mean square (RMS) deviation of the values from their arithmetic mean (Wikipedia)] compared to 81 sand bodies intersected by well NBU-222 (Table 6-4). Petrel, one more time, matches the characteristics of well NBU-222. 78, thus, is considered to be representative of the number of sand bodies penetrated by any well anywhere in the modeled volume. The total number of sand bodies intersected by a given number of wells in each well-spacing scenario is calculated in Table 6-5 and plotted in Figure 6-10. Refer to Appendix B for map views of the well distribution in each scenario. In a 1.35-ac (0.005 km2) well-spacing scenario, 636 wells were distributed around well NBU-222. The estimated total number of sand bodies intersected by all 636 wells is 49,608 (78 x 636) sand bodies. Therefore, when we drill at a 1.35-ac (0.005 km2) well spacing, we may penetrate all the sand bodies present, and 50% of the sand bodies (49,608 – 26,604) may be penetrated by more than one well. For the rest of the drilling scenarios, we would expect a linear relationship (y = 78x; Figure 6-11A) to relate the number of wells (x) in a drilling scenario to the number of sand bodies intersected (y). Consequently, as shown in Table 6-5, the wells in the other well-spacing scenarios penetrated a cumulative number of sand bodies less than 26,604 bodies. Therefore, we could not infer the number of sand bodies (if any) penetrated by more than one well in each of these scenarios. The relationship relating the number of wells (x) in a drilling scenario to the percentage of sand bodies intersected (y) is y = 0.41 x0.89 (Figure 6-11B). Table 6-5 shows the percentages in each drilling scenario.

237

Figure 6 - 10: A plot to show the number of wells and the total number of sand bodies intersected and their percentages in different infill-drilling scenarios. A total of 26,604 sand bodies were reported in the Petrel model. Using 78 as the average number of sand bodies intersected by a well in the modeled area, the 636 wells in the 1.35-ac well-spacing scenario would intersect a total of 49,608 sand bodies.

238

Table 6 - 5: Summary of the well-spacing scenarios considered showing the numbers of wells and sand bodies intersected in each.

Figure 6 - 11: Relationships of the number of wells in a drilling scenario to (A) the number of sand bodies intersected and (B) the percentage of the sand bodies intersected.

239

6.5. Discussion The Petrel facies model is a geologic model built in fluvial sections of the Mesaverde Group to present a visualization of the subsurface sand distribution. Although no uncertainty analysis was conducted, the object-based modeling utilized “available data” (FMI facies interpretation, sand dimensions), was “deterministic only at well locations, created realistic reservoir body geometries, and represented statistical variability … of geological objects” (Martinius and Naess, 2005). Petrel showed that it can honor the input data (facies proportions, N/G, and number of intersected sand bodies) with only one control point (well NBU-222) in the model constrained by 34 other wells. We could meet the objective of the modeling exercise and determine the percentages of sand bodies intersected in the various drilling scenarios. With the dimensions, orientations, and elliptical geometry of the channel-sand deposits, we could show that drilling efficiency intersecting more sand bodies and/or geobodies increases with closely spaced wells (Table 6-5). We obtained that 12.0%, 23.7%, and 49.5% of the total sand bodies are intersected in 20-ac (0.08 km2), 10-ac (0.04 km2), and 5-ac (0.02 km2) spacing scenarios, respectively. Therefore, drilling efficiency seems to be most significant at the 10-ac (0.04 km2) and 5-ac (0.02 km2) spacings. With the sand areas reduced in half in our model, we can see that our percentages and Panjaitan’s (2006) are remarkably different, especially in large well-spacing scenarios. Panjaitan (2006) reported percentages of 48%, 37%, 25%, and 20% of total sand bodies intersected by 10ac (0.04 km2), 20-ac (0.08 km2), 40-ac (0.17 km2), and 80-ac (0.32 km2) well-spacing scenarios, respectively. The number of sand bodies intersected by the 1.35-ac (0.005 km2) well-spacing scenario doubles the total number of sand bodies present in the modeled area (Table 6-5). A 1.35-ac (0.005 km2) well-spacing scenario, thus, is the approach to be followed if all the sand bodies present are to be intersected by one well or more. The positive aspect (intersecting all the sand bodies) of such a drilling scenario may at the same time be a drawback for tight/close well-spacing scenarios. In a system where the sand bodies are of low permeability with no proven hydraulic connectivity among them, depleting reservoir pressure becomes an issue to address. In such a case, penetrating the same sand body

240

with more than one well can lead to a faster pressure depletion and a shorter wellproduction period. Therefore, a 1.35-ac (0.02 km2) drilling scenario should not be attempted or implemented at a large scale before pressure testing and planning of the production life of each drilled well. Also, with a 1.35-ac (0.02 km2), the cost and logistics of drilling become an issue. Local regulations and the field reserves may not justify such fine-spacing drilling operations. Therefore, the main criteria that dictate which drilling scenario is the optimum scenario are the sand-body dimensions, cost of drilling, and estimated reserves in the field. Consequently, drilling at a 1.35-ac (0.02 km2) may not be possible and the 10-acre (0.04 km2) and, more favorably, 5-ac (0.02 km2) spacing scenarios become the optimum achievable or realistic scenarios. The section of the Mesaverde Group modeled in well NBU-222 contains sand in 48% of its thickness (Figure 6-12). We could compare our N/G results to other modeling studies. Larue and Hovadik (2006) showed that a high N/G (>30%) imposes a high stratigraphic connectivity among the sand bodies. Therefore, no remarkable increase in the sand volumes (Appendix B), connected to the wells of the various well-spacing scenarios, was seen. With no permeability or porosity data included in the model, we could not study or infer the hydraulic connectivity pattern, nature, or extent in the model.

241

Figure 6 - 12: Facies proportions in MVU and MVL of well NBU-222.

242

CHAPTER SEVEN

CONCLUSIONS AND RECOMMENDATIONS

7.1. Conclusions This study can be divided into three parts: a structural part where the principal stress orientations were determined, a stratigraphic part where the environments of deposition of the various sandstone packages were determined, and a geological modeling part where a facies model was created to reflect the subsurface sandstone distribution. The main conclusions are: 1- The main structural elements observed in the well and analyzed for stress orientation were open natural fractures, drilling-induced fractures, and borehole breakouts. The open natural fracture and drilling-induced fracture sets, on one hand, and the borehole breakouts, on the other, determined the orientations of the maximum stress (SHmax) and the minimum stress (SHmin), respectively. The mean SHmax and SHmin obtained from the 4 wells were found to trend 103 and 17 degrees from north, respectively. The results were similar to previous studies done in the area as well as the stress regional trend. 2- For all wells combined, channel sand, crevasse splays, and shoreface sand form 33%, <5%, and 10%, respectively, of the total logged intervals. A deviation of 4% to 15% from the average was seen in individual wells due to natural variation and the fact that the wells do not penetrate the same Mesaverde Group and Castlegate Formation intervals. Coal, lagoonal and washover sands existed in minor amounts. 3- Stratigraphic analysis shows that the logged intervals of the Mesaverde Group contain sand and shale in almost equal proportions. The average net-to-gross ratio in the studied Mesaverde Group and the Castlegate Sandstone was found to be 0.55.

243

4- The Upper Mesaverde Group was deposited in a fluvial environment, the Lower Mesaverde Group contained deposits from fluvial, shoreface and transition environments, and the Castlegate Sandstone was deposited in a shallow marine environment. 5- Geological modeling constitutes a link between image-log interpretation and facies modeling; an approach usually not adopted by either service companies or oil companies. With the facies interpretation, we obtained the vertical distribution of the various facies to constrain and control the geological model. 6- As expected, with more wells drilled, more subsurface sand bodies are intersected. More precisely, we present a relationship between the number of wells (x) in a drilling scenario and the percentage of sand bodies intersected (y), where y = 0.41 x0.89. 7- Drilling at 41-ac (0.17 km2) or larger can make an operator miss more than 90% of the sand bodies in a field. 8- With wells drilled at 1.35-ac (0.005 km2), 50% of the sand bodies could be penetrated by more than one well. 9- The optimum realistic drilling scenario in a similar fluvial system would be the 10-ac (0.04 km2) or 5-ac (0.02 km2) spacing scenario.

7.2. Recommendations The results from structural interpretation could only specify the direction along which hydraulic fracturing should form (parallel to SHmax). However, were the difference in magnitude between the maximum and minimum stresses studied from the image logs, obtained from previous research, or used as reported from operating parties in the fields, more recommendations could have been possible to make on the type and specifications of the hydraulic fracturing jobs. This could have enriched the research by providing a better understanding or recommendations for field development and well completion. An ideal approach to study subsurface sand could have been integrating studies done on outcropping sand from the various sandstone formations and calibrating logs with available cores. Seeing the continuity of the various Mesaverde Group sandstone 244

formations at the surface could have improved the understanding of the subsurface occurrence of the sand bodies. No personal work was done on outcrop analogs. Literature and personal communications were used, instead, to understand the subsurface sand architecture. Integrating facies interpretation from image logs into facies modeling is recommended as it is a tool to restrict and control a geological model at specific well locations. A geologist, therefore, can introduce into the model a criterion to control the geometry, dimensions, and orientations of the sand bodies forcing the modeling application (Petrel) to honor an additional set of data other than the conventional wireline measurements. Results from Chapter 6 showed that the more we drill the more sand bodies can be accessible, but not necessarily more connected sand volumes. In a system with very low permeability reaching the orders of mD and µD, the hydraulic connectivity and flow are important parameters to determine. Therefore, including permeability and porosity data (values/curves/contour maps) as input data could have enabled us to highlight the nature of the hydraulic connectivity in the modeled area. Comparing structural features (fracture length, fracture porosity, and fracture density) and stratigraphic features (sand dimensions, geometries, and distribution) to available production data and results reported in the wells/fields would have been an elegant step to complete this research. Production data could have been used to relate production rates to fracture density and/or trace lengths in the sand. Also, production data could have shed some light on whether the sand geometries and dimensions used in Petrel are reasonable. Perhaps, a relationship between hydraulic connectivity and stratigraphic connectivity could have been understood or reached. The production data does help estimate the area/volume affected by production around a well. This would have reflected whether the sand dimensions and distribution specified in the model support the volume drained by production. Also, a better emphasis on the importance of studying hydraulic connectivity in low-permeability sand bodies would have been showed. All in all, although no approach or outline of how the production data could have been used (as the step was not conducted), nor conclusions drawn, the data could have helped fine-tune the geological model and highlight ways to improve it. 245

REFERENCES

Anderson, D. S., 2005, Architecture of crevasse splays and point-bar bodies of the nonmarine Iles Formation north of Rangely, Colorado: Implications for reservoir description: The Mountain Geologist, v. 42, No. 3, p. 109-122. Babcock, E. A., 1978, Measurement of subsurface fractures from dipmeter logs: AAPG Bulletin, v. 62, p. 1111-1126. Bally, A. W., Scotese, C. R., and Ross, M. L., 1989, North America: Plate-tectonic setting and tectonic elements, in A. W. Bally, and A. R. Palmer, eds., The Geology of North America – An overview: Boulder, Colorado, Geological Society of America, the Geology of North America, v. A., p. 1-17. Blakey, R. C., 2003, Paleogeography through geologic time: GIF image from http://jan.ucc.nau.edu/~rcb7/crepaleo.html. Bridge, J. S., and Mackey, D., 1993, A theoretical study of fluvial sandstone body dimension, in S. S. Flint, and I. D. Bryant, eds., Geological modeling of hydrocarbon reservoirs: International Association of Sedimentologists Special Publication 15, p. 213-236. Cole, R. D., and Cumella, S. P., 2005, Sand body architecture in the Lower Williams Fork Formation (Upper Cretaceous), Coal Canyon, Colorado, with comparison on the Piceance basin subsurface: The Mountain Geologist, v. 42, No. 3, p. 85107. Cole, R., and Cumella, S. P., 2003, Stratigraphic architecture and reservoir characteristics of the Mesaverde Group, southern Piceance basin, Colorado, in K. M. Peterson, T. M. Olson, and D. S. Anderson, eds., Piceance Basin 2003 Guidebook: Chapter 18. Rocky Mountain Association of Geologists and the American Institute of Professional Geologists, Denver, Colorado, 449 p. Cumella, S. P. and Ostby, D. B., 2003, Geology of the basin-centered gas accumulation, Piceance basin, Colorado, in K. M. Peterson, T. M. Olson, and D. S. Anderson, eds., Piceance basin 2003 guidebook: Chapter 10. Rocky Mountain Association of Geologists and the American Institute of Professional Geologists, Denver, Colorado, 449 p. Cuzella, J. J., and Stancel, S. G., 2006, Greater Natural Buttes field, Uinta basin, Utah: The Mountain Geologist, v. 43, No. 3, p. 213-218.

246

Donselaar, M. E., and Schmidt, J. M., 2005, Integration of outcrop and borehole image logs for high-resolution facies interpretation: Example from a fluvial fan in the Ebro basin, Spain: Sedimentology, 52, p. 1021-142. Enchanted Learning accessed on June 7th, 2006 from: http://www.enchantedlearning.com/usa/statesbw/mountain/mountain.shtml Friedman, G. M., and Sanders, J. E., 1978, Principles of Sedimentology: John Wiley & Sons, New York, USA, 792 p. Galloway, W. E., and Hobday, D. K., 1996, Terrigenous clastic depositional systems: Springer-Verlag Berlin Heidelberg, Italy, 489 p. German, Q. A., 2006, Analysis of fluvial sandstone-body characteristics and architecture in a high net-to-gross system: Upper Williams Fork Formation, Plateau Creek Canyon, Piceance basin, Colorado: Master’s thesis, University of Colorado, Boulder, Colorado, 151 p. Gilreath, J. A., 1987, Dipmeter interpretation rules: The Technical Review, Schlumberger Ltd., Houston, Texas, USA. Grace, L. M., and Newberry, B. M., 1998, Geological application of dipmeter and borehole electrical images: Unpublished Short Course Notes, Schlumberger Oilfield Services, Texas, USA. Hallam, A., 1981, Facies interpretation and the stratigraphic record. W. H. Freeman and Company, Oxford, 291 p. Henshaw, E., 2005, Petrel-based 3-D structural and facies modeling of the Riverbend area, Uinta basin, Utah, US land west. Unpublished internal Schlumberger report, USA. Hettinger, R. D., and Kirschbaum, M. A., 2003, Stratigraphy of the Upper Cretaceous Mancos Shale (upper part) and Mesaverde Group in the southern part of the Uinta and Piceance basins, Utah and Colorado: USGS Digital Data Series DDS-69-B, Chapter 12. Hintze, L. F., 1988, Geologic history of Utah: Brigham Young University Geology studies. Special publication No. 7, Brigham Young University, Utah, 202 p. Hoffman, P. F., 1989, Precambrian geology and tectonic history of North America, in A. W. Bally, and A. R. Palmer, eds., The Geology of North America – An overview: Boulder, Colorado, Geological Society of America, the Geology of North America, v. A., p. 447-512.

247

Hurley, N. F., 2004, Borehole images, in G. Asquith, and D. Krygowski, Basic well log analysis: AAPG Methods in Exploration 16, p. 151-163. Hurley, N. F., 1994, Recognition of faults, unconformities, and sequence boundaries using cumulative dip plots: AAPG Bulletin, v. 78, no. 8, p. 1173-1185. Johnson, R. C., 2003, Depositional framework of the Upper Cretaceous Mancos Shale and the lower part of the Upper Cretaceous Mesaverde Group, western Colorado and eastern Utah: USGS Digital Data Series DDS-69-B, Chapter 10, Denver, Colorado, 24 p. Johnson, R. C., 1993, Surface vitrinite reflectance study of the Uinta and Piceance basins and adjacent areas, eastern Utah and western Colorado – implications for the development of Laramide basins and uplifts: USGS Bulletin 1787-DD, Denver, Colorado, 38 p. Johnson, R. C., and Roberts, S. B., 2003, The Mesaverde total petroleum system, UintaPiceance province, Utah and Colorado: USGS Digital Data Series DDS-69-B, Chapter 7, Denver, Colorado, 65 p. Johnson, S. Y., 1992, Phanerozoic evolution of sedimentary basins in the Uinta-Piceance basin region, northwestern Colorado and northeastern Utah: U.S. Geological Survey Bulletin 1787-FF, Denver, Colorado, 38 p. Knight, C. N., 1999, Structural and stratigraphic controls on Mesaverde reservoir performance: North La Barge field, Sublette County, Wyoming: Unpublished Ph.D. Dissertation, Colorado School of Mines, Golden, Colorado, 217 p. Koepsell, R., Cumella, S. P., and Uhl, D., 2003, Applications of bore hole images in the Piceance basin, in K. M. Peterson, T. M. Olson, and D. S. Anderson, eds., Piceance Basin Guidebook: Chapter 14. Rocky Mountain Association of Geologists and the American Institute of Professional Geologists, Denver, Colorado, 449 p. Larue, D. K., and Hovadik, J., 2006, Connectivity of channelized reservoirs: a modelling approach, Petroleum Geoscience, v. 12, p. 291-308. Lawton, T. F., 1986, Fluvial systems of the Upper Cretaceous Mesaverde Group and Paleocene North Horn Formation, central Utah – A record of transition from thinskinned to thick-skinned deformation in the foreland region, in J. A. Peterson, ed., Paleotectonics and sedimentation in the Rocky Mountain Region, United States: AAPG Memoir 41, p. 423-442. Longman, M. W., and Koepsell, R. J., 2005, Defining and characterizing Mesaverde and Mancos sandstone reservoirs based on interpretation of the image logs, Eastern Uinta Basin: Utah Geological Survey, Utah, USA, Open-file report 458, 122 p. 248

Lorenz, J. C., 2003, Fracture systems in the Piceance basin, overview and comparison with fractures in the San Juan and Green River basins, in K. M. Peterson, T. M. Olson, and D. S. Anderson, eds., Piceance Basin Guidebook: Chapter 5. Rocky Mountain Association of Geologists and the American Institute of Professional Geologists, Denver, Colorado, 449 p. Lorenz, J. C., and Finley, S. J., 1991, Regional fractures II: fracturing of Mesaverde reservoirs in the Piceance basin, CO: AAPG Bulletin, v. 75, p. 1738-1757. Lorenz, J. C., Heinze, D. M., Clark, J. A., and Searls, C. A., 1985, Determination of widths of meander-belt sandstone reservoirs from vertical downhole data, Mesaverde Group, Piceance basin, Colorado: AAPG Bulletin, v. 69, p. 710-721. Luthi, S. M., and Souhaite, P., 1990, Fracture apertures from electrical borehole scans: Geophysics, v. 55, p. 821-833. Luthi, S. M., 2001, Geological well logs: Springer-Verlag, Berlin, 373 p. Martinius, A. W., and Næss, A., 2005, Uncertainty analysis of fluvial outcrop data for stochastic reservoir modelling, Petroleum Geoscience, v. 11, p. 203-214. Oldow, J. S., Bally, A. W., Avé Lallemant, H. G., and Leeman, W. P., 1989, Phanerozoic evolution of the North American Cordillera; United States and Canada, in A. W. Bally, and A. R. Palmer, eds., The Geology of North America – An overview: Boulder, Colorado, Geological Society of America, the Geology of North America, v. A., p. 139-232. Osmond, J. C., 1965, Geologic history of site of Uinta basin, Utah: AAPG Bulletin, v. 49, p. 1957-1973. Osmond, J. C., Locke, R., Dille, A. C., Praetorious, W., and Wilkins, J. G., 1968, Natural gas in Uinta Basin, Utah: AAPG Special Volumes, v. A007, p. 174-198. Panjaitan, H., 2006, Sand-body dimensions in outcrop and subsurface, lower Williams Fork Formation, Piceance basin, Colorado: unpublished Master’s thesis, Colorado School of Mines, Golden, Colorado, 170 p. Pitman, J. K., Franczyk, K. J., and Anders, D. E., 1987, Marine and nonmarine gasbearing rocks in the Upper Cretaceous Blackhawk and Neslen Formations, Eastern Uinta basin, Utah: Sedimentology, diagenesis, and source rock potential: AAPG Bulletin, v. 17, p. 76-94. Pranter, M. J., Reza, Z. A., and Cole, R. D., in prep., Reservoir-scale connectivity of fluvial sandstones: an example from the Williams Fork Formation, Piceance Basin, Colorado, U.S.A.

249

Patterson, P. E., Kronmueller, K., and Davies, T. D., 2003, Sequence stratigraphy of the Mesaverde Group and Ohio Creek Conglomerate, northern Piceance basin, Colorado, in K. M. Peterson, T. M. Olson, and D. S. Anderson, eds., Piceance basin 2003 guidebook: Chapter 7. Rocky Mountain Association of Geologists and the American Institute of Professional Geologists, Denver, Colorado, 449 p. Pitman, J. K, Franczyk, K. J., and Anders, D. E., 1987, Marine and nonmarine gasbearing rocks in Upper Cretaceous Blackhawk and Neslen Formations, Eastern Uinta Basin, Utah: Sedimentology, diagenesis, and source rock potential: AAPG Bulletin, v. 17, p. 76-94. Plumb, R. A., and Hickman, S. H., 1985, Stress-induced borehole elongation: a comparison between the four-arm dipmeter and the borehole televiewer in the Auburn geothermal well: Journal of Geophysical Research, v. 90, p. 5513-5521. Prothero, D. R., 1990, Interpreting the stratigraphic record: W. H. Freeman and Company, New York, 410 p. Raisz, E., 1972, Physiographic provinces in the Rocky Mountain region and landforms of western United States, in W. W. Mallory, ed., Geologic Atlas of the Rocky Mountains region: Rocky Mountain Association of Geologists, 331 p. Rider, M., 1996 (2nd edition), The geological interpretation of well logs. Whittles Publishing, Scotland, 280 p. Rowley, P. D., Hansen, W. R., Tweto, O., and Carrara, P. E., 1985, Geologic map of the Vernal 1º x 2º quadrangle, Colorado, Utah, and Wyoming. pdf file accessed on February 26th, 2007 from: http://geology.utah.gov/maps/geomap/1x2/index.htm Schlumberger, 2002, FMI: borehole geology, geomechanics and 3-D reservoir modeling, 2006 from: pdf files accessed on June 1st, http://www.slb.com/media/services/evaluation/geology/fmi.pdf Scholle, P. A., and Spearing, D., (eds.), 1982, Sandstone depositional environments: AAPG Memoir 31, 418 p. Selley, R. C., 1978, Ancient sedimentary environments, 2nd edn., Chapman and Hall, London, 287 p. Serra, O., 1989, Formation MicroScanner image interpretation: Houston, Texas, Schlumberger Educational Services, 117 p. Serra, O., 1985, Sedimentary environments from Wireline logs. Schlumberger, 211 p.

250

Shawa, M. S. (ed.), 1979, Use of sedimentary structures for recognition of clastic environments (second edition), Canadian Society of Petroleum Geologists Publication, 66 p. Spencer, C. W., 1987, Hydrocarbon generation as a mechanism for overpressuring in Rocky Mountain region: AAPG Bulletin, v. 71, p. 368-388. Stokes, W. L., 1986, Geology of Utah: Utah Museum of Natural History and Utah Geological and Mineral Survey, Salt Lake City, Utah, 307 p. Stokes, W. L., 1977, Physiographic subdivisions of Utah: Utah Geological and Mineral Survey Map 43, scale 1:2,500,000. Zoback, M. D., Moos, D., Mastin, L., and Anderson, R. N., 1985, Well bore breakouts and in situ stress: Journal of Geophysical Research, v. 90, p. 5523-5530. Zoback, M. L., and Zoback, M. D., 1989, Tectonic stress fields of the continental United States, in L. C. Pakiser, and W. D. Mooney, Geophysical framework of the continental United States: Boulder, CO, Geological Society of America, Memoir 172, p. 523-539.

251