Abstract There are a variety of pore types in unconventional resource mudstones and shales. The currently preferred method by geologists and petrophysicists is to examine and analyze these mudstones and shales by argonion and focused ion-beam milling to produce an ultrasmooth surface, coupled with observation under the field emission scanning electron microscope (FESEM). Potential issues with Ar-ion milled/FESEM preparation and imaging include (1) the small size of sample cubes for upscaling, (2) loss of structural fabric during the milling–imaging process, (3) fewer non-inorganic pore types observed than when observed with an unpolished surface, (4) analog use of pores from one shale to another, although the pore types and composition might differ, and (5) the creation of potential artifacts related to desiccation and rock expansion because of core retrieval and sample preparation. Conventional FESEM images obtained from freshly broken surfaces reveal much more textural detail than those obtained from ion-milled (polished) surfaces. Although conventional FESEM methodology may share some of the same limitations as Ar-ion beam-milled/FESEM technology, FESEM methodology should not be overlooked because it provides a more cost-effective and potentially more accurate analysis for estimating porosity and determining pore types and their distribution in shales. Comparison of FESEM images from ion-milled and fresh, non-ion-milled surfaces reveals that organic matter and internal organoporosity are best viewed on ion-milled surfaces, but shale microfabric and non-organoporosity is best viewed under non-milled surfaces. Complete FESEM imagery for shale characterization should include both types of analyses.

Abstract The slope morphologies of the Magdalena deepwater fan exhibit a series of channel-levee complexes (CLCs), recording the evolution of the Magdalena delta. Detailed morphological analysis of the seafloor expression of the channels and their lateral relationship allows the reconstruction of the history of Pleistocene fan development. The Magdalena deepwater fan was deposited on the northern offshore Colombia accretionary wedge (Caribbean Sea), initiated during the late Miocene. Fan evolution is closely related to the Magdalena River delta migration and reflects control by tectonic processes occurring from Pliocene to present. Major delta shifts toward the southwest (Canal del Dique) and northeast (Cienaga de Santa Marta region) create a submarine fan that migrated with the river, becoming younger toward the southwest. The main fan was abandoned during the Holocene, focusing deposition on the Barranquilla region to the northeast with modern active sedimentation. The depositional processes in the active fan area are mainly dominated by turbidity currents, alternating with slumps/debris flows that generated large mass transport deposits (MTDs). Eight river delta phases were identified, linked to the onshore geology and their corresponded submarine fan expression, which is characterized by the presence of CLCs and MTDs. Seven CLCs were studied using multi-beam bathymetry and seismic profiles. The CLCs showed a big variation of sinuosity and gradient throughout the slope. The higher sinuosity values were encountered at areas of high gradients, suggesting that the channels attempt to reestablish its equilibrium profile by increasing sinuosity as a response of changes in the slope. Highly sinuous channels in the western fan suggest that sinuosity changes are controlled by changes on the slope associated with the deformation of the fold-and-thrust belt along the margin. In addition, channel’s forced migration, avulsions, convex-up profiles, and the presence of knickpoints (KPs) suggest ongoing deformation during western CLC deposition. Conversely, the northeastern section of the fan exhibits channel thalweg profiles with lower sinuosity values at deeper depths. Convex-up thalweg profiles in this area may represent disequilibrium profiles or post-abandonment deformation. Older CLCs are highly affected by degradational processes after the abandonment of the systems, increasing channel width and modifying levee walls. These processes should be considered when evaluating dimensions of buried deposits and reservoir quality prediction. A sequence of KPs in the western fan seems to connect sediment flows from the shelf break downslope through a series of steps in the slope, culminating with lobate unconfined deposits. Upstream KP migration in slope steps as a response to deformation may represent a key process to explain the initiation of deepwater channel systems on the Magdalena Fan, as well as channel systems deposited on other tectonically active basins. This study provides new understanding of the processes involved in the Magdalena deepwater fan and implications for channel systems characterization in areas with active deformation during deposition.

Abstract This paper describes and illustrates features of shales and mudstones at the nanometer and micrometer scales using standard scanning electron microscopy (SEM) and field emission scanning electron microscopy (FE-SEM) techniques. Microfabric observations at these scales not only provide insights into sedimentary and postdepositional processes but also offer evidence useful in understanding storage and primary migration patterns in unconventional shales. The images illustrated are suggested as references to guide future shale studies related to shale porosity and permeability. Examples are provided of various shales (Barnett, Woodford, Eagle Ford, Rhinestreet, Fayetteville, and Marcellus). Microfabric and pore features illustrated include clay flakes related to open-network floccules and clay-alighed fabric, plus other features produced by nonclay minerals. Organic matter produced by zooplankton and algae (e.g., coccolithophores, Tasmanites) is described because it may form organic mucus that adheres to and helps aggregate clay flakes. Organic matter is also common within porous fecal pellets. Coccolithophores, sponge spicules, and foraminifera tests contain hollow internal chambers, which provide porosity and probably permeability, even when filled with clay minerals. Conventional SEM provides a rapid and relatively inexpensive way of evaluating pores and microfabrics in shales.

Abstract Based on recent studies of Barnett and Woodford gas shales in Texas and Oklahoma, a systematic characterization workflow has been developed that incorporates lithostratigraphy and sequence stratigraphy, geochemistry, petrophysics, geomechanics, well log, and three-dimensional (3-D) seismic analysis. The workflow encompasses a variety of analytical techniques at a variety of geologic scales. It is designed as an aid in identifying the potentially best reservoir, source, and seal facies for targeted horizontal drilling. Not all of the techniques discussed in this chapter have yet been perfected, and cautionary notes are provided where appropriate. Rock characterization includes (1) lithofacies identification from core based on fabric and mineralogic analyses (and chemical if possible); (2) scanning electron microscopy to identify nanofabric and microfabric, potential gas migration pathways, and porosity types/distribution; (3) determination of lithofacies stacking patterns; (4) geochemical analysis for source rock potential and for paleoenvironmental indicators; and (5) geomechanical properties for determining the fracture potential of lithofacies. Well-log characterization includes (1) core-to-log calibration that is particularly critical with these finely laminated rocks; (2) calibration of lithofacies and lithofacies stacking patterns to well-log motifs (referred to as gamma-ray patterns or GRPs in this chapter); (3) identification and regional to local mapping of lithofacies and GRPs from uncored vertical wells; (4) relating lithofacies to petrophysical, geochemical, and geomechanical properties and mapping these properties. Three-dimensional seismic characterization includes (1) structural and stratigraphic mapping using seismic attributes, (2) calibrating seismic characteristics to lithofacies and GRPs for seismic mapping purposes, and (3) determining and mapping petrophysical properties using seismic inversion modeling. Integrating these techniques into a 3-D geocellular model allows for documenting and understanding the fine-scale stratigraphy of shales and provides an aid to improved horizontal well placement. Although the workflow presented in this chapter was developed using only two productive gas shales, we consider it to be more generically applicable.

Abstract An outcrop-behind outcrop study was conducted in and adjacent to a 300 × 100 × 16 m (980 × 330 × 50 ft) quarry of the gas-producing Woodford Shale to structurally/stratigraphically characterize it from the pore to subregional scales using a variety of techniques. Strata around quarry walls were described and correlated to a 64 m (210 ft) long continuous core drilled 150 m (500 ft) back from the quarry wall and almost to the Woodford-Hunton unconformity. Borehole logs obtained include neutron and density porosity (NPHI and DPHI) logs, and logs from Elemental Capture Spectroscopy (ECS™), Combinable Magnetic Resonance (CMR-Plus™), Fullbore Formation MicroImager (FMI™), and sonic scanner (Modular Sonic Imaging Platform, or MSIP™)—all manufactured by Schlumberger. The strata around the quarry are horizontally bedded. Borehole logs were used to identify a basic threefold subdivision into an upper relatively porous quartzose interval; a middle, more clay-rich, and less porous interval; and a lower interval of intermediate quartz-clay content. These intervals correspond to the informally named upper, middle, and lower Woodford. Detailed core and quarry wall description revealed several types of finely laminated lithofacies, with varying amounts of total organic carbon (TOC). The FMI log revealed a much greater degree of variability in laminations than can be readily seen with the naked eye. Organic geochemistry and biomarkers are closely tied to these lithofacies and record cyclic variations in oxic-anoxic depositional environments, which correspond to relative sea level fall-rise cycles. At the scanning electron microscopy scale, microfractures and microchannels are common and provide tortuous pathways for gas (and oil) migration through the shales. Based on FMI and core analysis, fracture density is much greater in the upper quartzose lithofacies than in the lower, more clay-rich lithofacies. A laser imaging detection and ranging (LIDAR) survey around the quarry walls documented two near-vertical fracture trends in the quartzose lithofacies: one striking N85°E with spacings of 1.2 m (4 ft) and the other striking N45°E related to the present stress field. The FMI analysis only imaged the latter fracture set. Both log-derived and laboratory-tested geomechanical property measurements documented a significant relationship between shale fabric (laminations and preferred clay-particle orientation) and rock strength, and a secondary relationship to mineral composition. Porosity and microfractures or microchannels also appear to influence rock strength. This integrated study has provided insight into the causal relations among Woodford properties at a variety of scales. In particular, a stratigraphic (vertical) segregation of lithofacies can be related to cyclic variations in depositional environments. The resulting stratified zones exhibit variations in their hydrocarbon source and reservoir (fracturable) potential. Such information and predictive capability can be valuable for improved targeted horizontal drilling into enriched source rock and/or readily fracturable reservoir rock in the Woodford and perhaps other gas shales.

Abstract The sequence-stratigraphic framework established for the subsurface Barnett Shale in the northern part of the Fort Worth Basin is helping to resolve the age, nature, and fill of karst features under the Barnett Shale in the southwestern part of the basin. The southwestern Fort Worth Basin is characterized by the absence of the Upper Ordovician Viola Limestone and Simpson Group, which separate the lower Barnett Shale from the underlying Ordovician Ellenburger Group, as well as the Forestburg Limestone, which separates the upper and lower Barnett Shale to the north. Consequently, the undifferentiated Barnett Shale unconformably overlies the water-bearing Ellenburger Group and results in a higher risk of water encroachment during stimulation and production of Barnett gas wells. Recent work indicates that Barnett Shale parasequence sets dominated by phosphatic and siliceous shale lithofacies are more organic rich and possibly more gas prone than other Barnett lithofacies. Moreover, the quartz- and carbonate-rich lithofacies are brittle and appear to respond more favorably to hydrofracture stimulation and the facies with high amounts of clay may serve as a possible barrier for fracture propagation because of ductile behavior. Thus, the ability to locate and map these parasequence sets was a particularly important part of this study for aiding in reservoir characterization. Analysis of three-dimensional seismic data southwest of the core area of the Newark East field demonstrates the ability to identify and map Barnett parasequence sets previously defined from core and logs in the more northerly part of the basin. In addition, high-resolution seismic images of the karsted Ellenburger Group unconformity surface reveal a series of elongate, rectilinear, collapsed paleocave systems resulting from subaerial exposure and carbonate dissolution. These features appear to have shaped the unconformity surface and to have had a direct influence on the deposition and distribution of the overlying Barnett Shale parasequence sets. The parasequence sets are thicker over these collapsed features than in areas flanking the karst. The difference in thickness diminishes with each stratigraphically younger parasequence set, indicating focused infilling over the collapsed features caused by progressive reduction in accommodation space. Seismic analysis also reveals that the karst topography on the unconformity surface is related not only to local faulting caused by the paleocave collapse, but also to deep-seated northwest–southeast-trending faults that extend upward to the Ellenburger surface and sometimes into the overlying Barnett Shale, suggesting post-karst fault movement. Magnetic surveys over the area support the deeper origin of the fault pattern observed in the study area. In the Newark East field, the Viola Limestone and Simpson Group form a fracture barrier for the overlying Barnett Shale. Their absence to the southwest presents a dilemma—whereas the Barnett Shale is thicker over this area, the lack of a fracture barrier risks water encroachment from the underlying Ellenburger Group. Understanding Ellenburger karst development and behavior and how fault and fracture systems are associated with these structures is critical for comprehending the distribution and depositional pattern of the Barnett Shale parasequence sets. Moreover, the seismic mapping and characterization of the different parasequence sets (ranging in lithofacies and rock property) would allow improvement in selecting horizontal targets and fracture stimulation of Barnett gas wells.

Abstract Although originally used for structural interpretation, borehole image and dipmeter logs are finding progressively more application to identifying sedimentary features in both exploration and development wells. The general consensus has been that to gain the most sedimentary information from image logs, they should be calibrated to rocks, particularly to corresponding cores. We describe five case studies that calibrate image logs not only to cores, but also to outcrops of analogous facies, using behind-outcrop wells. In particular, the behind-outcrop logging has proven very successful in the calibration process, particularly as it pertains to identifying features that provide indications of sedimentary features laterally away from the wellbore. The five case studies are from three deep-water (turbidite) rock sequences: Pennsylvanian Jackfork Group (Arkansas), Miocene Mt. Messenger Formation (New Zealand), and Cretaceous Dad Sandstone Member of the Lewis Shale (Wyoming). Behind-outcrop wells have been drilled, logged, and cored for each area. The results demonstrate the following: (1) Although calibrating image logs of deep-water strata to corresponding core is strongly recommended, it is not always essential because features illustrated in this chapter and other publications provide a partial catalog for characterization. (2) Interpretation of sedimentary environments and facies from an image log requires the identification of a group of features, which can be related to sedimentary processes, instead of a single feature. (3) Laminae-scale stratification features are commonly more readily seen on the image log than in corresponding core; however, some small-scale sedimentary structures may be difficult or even impossible to identify. (4) Dipmeter logs can be used to differentiate certain deep-water facies, even when from old wells, provided the structural dip is deleted from the data set. These findings provide a variety of applications to both exploration and development, for example, in the areas of volumetric and net sand calculation, prediction of reservoir trend and geometry, and the overall value of obtaining borehole image logs, which are less expensive to obtain than core.

Abstract The Cretaceous Lewis Shale of Southern Wyoming is an excellent outcrop example of a submarine fan deposited basinward of a coeval, prograding margin. A regional cross section constructed from outcrop and subsurface data reveal several large-scale attributes of the system. Regionally continuous, condensed sections in the Lewis Shale define southward-prograding clinoforms that are related to a shelf-slope-basin physiography during deposition. The condensed sections form the boundaries to fourth-order stratigraphic cycles/parasequences. The average height of the clinoforms is ~400 m (~1300 ft), which is interpreted to reflect the minimum water depth during deposition. Strata with 50% or more sandstone are located in fluvial-deltaic strata on the topset of the clinoforms and submarine-fan strata are located on the bottomset of the clinoforms. Slope strata on the foresets of clinoforms contain 15–20% sandstone. Although the sandiest strata are located in shelf and base-of-slope strata, the depocenter of each fourth-order cycle is consistently located in muddy slope strata.

Abstract The Cretaceous Lewis Shale-Fox Hills-Lance Formations in southern Wyoming ( Figure 1 ) were deposited during the final transgressive-regressive cycle of the Cretaceous western interior seaway ( Figure 2 ). These formations comprise a third-order highstand systems tract ( Figure 3 ) composed of several fourth-order lowstand-highstand cycles. One of these fourth-order cycles is well exposed in three outcrops — called Spine 1, Spine 2, and Rattlesnake Ridge — over a distance of 3.2 km (2 mi) ( Figure 4 ). The strata within these outcrops are important for two reasons: 1) they provide good outcrop exposures of the Lewis Shale in an area of active exploration for gas, and 2) the Dad Sandstone Member of the Lewis Shale here is considered to be an excellent, scaled analog of delta-fed, mud-dominated, progradational, turbidite systems that are hydrocarbon-productive in areas such as the Gulf of Mexico and offshore west Africa.

Abstract A 3-D geological model was constructed from a 3-D outcrop for reservoir flow simulation that can address the effects of small-scale (subseismic), interwell heterogeneities on production in analog deep-water oil and gas reservoirs. The dimensions of the Hollywood Quarry, Arkansas (Figure 1) , are 380 x 250 x 25 m (1247 x 821 x 83 ft) ( Figures 2 , 3 ). The quarry exposes in 3-D the upper Jackfork Group turbidites, which ate often used as an outcrop analog for deep-water reservoirs in the Gulf of Mexico and elsewhere. A variety of turbidite facies are present: lenticular, channelized sandstones, pebbly sandstones, and conglomerates within shales (CI); laterally continuous, interbedded thin sandstones and shales (SI, S2); and thicker, laterally continuous shales (Ml, M2). Sandstone and shale beds are folded and cut by strike-slip faults with a vertical component. These combinations of structural elements and facies have resulted in a stratigraphic interval that is highly compartmentalized, both horizontally and vertically. The quarry is used here as an analog to a variety of subsurface reservoir types. Techniques used to characterize the quarry include behind-outcrop coring, outcrop gamma-ray (GR) logging, measured stratigraphic sections, sequential photography of the quarry walls, Digital Orthophoto-Quadrangle (DOQ) mapping, Ground Penetrating Radar (GPR), Global Positioning System (GPS), shallow, high-resolution seismic reflection, and GPS laser-gun positioning of geologic features in 3-D space. The west wall has been quarried back within 0.5 m (1.6 ft) of the first inline of an earlier 3-D GPR survey and coring operation. The

Abstract The growth in estimated ultimate recovery ( EUR ) of oil and gas fields over the course of their development has been recognized as a significant contributor to hydrocarbon supply, both in the United States and abroad. Data on changes in EUR have been examined for oil and gas fields discovered on the modern shelf of the Gulf of Mexico, in order to empirically determine the possible causes of these changes. Using a semilog regression model of EUR as a function of years since discovery, from 1975 through 2002, roughly half of fields in the study area grew and the balance either shrank or remained statistically unchanged. Fields that grew were typically large discoveries to start and the volumes by which they grew were log normally distributed. The fields making the largest contributions to aggregate growth typically had at least 20 reservoirs over at least 5,000 feet of charged section, which was deposited in generally prograda-tional environments at sediment accumulation rates between 500 and 2,500 feet per million years. The principal mechanism of field growth in the study area was through the discovery of new reservoirs. In the fields having the largest growth, these discoveries occurred in cycles based on stratigraphic interval. Within each cycle, the largest reservoirs were discovered early and the size of reservoir discoveries declined exponentially. Up to four major stratigraphically based cycles were observed; generally, but not always, each subsequent cycle added a smaller volume to EUR than those that preceded it. A secondary source of growth arises through the combined effects of recognizing an increased volume of reservoir rock containing reserves and improvement in recovery factors. The contributions of these mechanisms have been examined through analysis of single-reservoir fields and growth in fields after their last new reservoir discovery. Field growth is tied to the economic conditions surrounding oil and gas production. From the mid-1970s through mid-1980s, during a period of rising and high prices, large increases in oil and gas reserves were gained through new field discoveries, discovery of new reservoirs within fields and, to a lesser extent, positive reservoir volume revisions and increases in recovery factors. Price collapses in 1986 and again in 1998 are both reflected in reductions in field growth and actually declines in aggregate EUR . Although a short time series, EUR growth between the beginning of the current price recovery in 1998 and 2002 indicates that supply of new oil and gas in existing fields is becoming more inelastic. This is most probably due to two factors: depletion of the growth potential of old, very large fields; and because of the progressive decline in the sizes of new field discoveries and the high correlation between size and growth, as newer finds have smaller growth potential.

Abstract Schiehallion is a two billion barrel deepwater clastic reservoir, situated on the Atlantic margin of the UKCS , one of the world’s most hostile environments for hydrocarbon production. The field has been developed via subsea wells tied back to an FPSO , and is one of the first developments of its kind anywhere in the world. The field may be characterized as high productivity but low energy and, as a consequence, water injection is essential to maintaining production. However, the reservoir is channelized, faulted, and has varying degrees of connectivity between the compartments, so that a good understanding of these factors is necessary to optimize the water injection distribution. Our understanding of the ‘plumbing’, or connectivity between the wells, has evolved and matured over time, using a wide range of different data types, from the initial extended well test, through RFT’s , pressure transient analyses, interference testing, PLT’s , tracer and geochemical sampling, to bi-annual 4D seismic surveys using increasingly sophisticated processing and interpretation. Much of this understanding has been incorporated in a 3D model, which uses object modeling and seismic conditioning to represent the sand distribution. Potential barriers to flow are identified from seismic coherency analysis, and the strengths of these barriers have been used as the main history matching parameters. A key learning has been that all data needs to be interpreted with great care, and it is essential to integrate several data types in order to obtain reliable conclusions. The paper gives examples of data which has been invaluable, as well as examples where the data is ambiguous or misleading.

Abstract The Ormen Lange gas field, discovered in 1997 (Hydro operated License PL209) in 1000 m (3,281 ft) water depth and covering an area of ca. 350 km 2 (217 mi 2 ) was further appraised by four wells prior to development approval in April, 2004. The partnership, Hydro (development operator), Norske Shell (production operator), Petoro, Statoil, ExxonMobil, and Dong, had a planned production-start in October, 2007, from 8 of 28 possible production wells in a staged development using four subsea templates. The development faced a number of challenges; rough seabed topography, subzero sea bottom temperatures, harsh ocean conditions and a change of operatorship at production start-up. Reservoir characterization of the areally limited, but intensely faulted turbidite reservoir has formed an integral part of the work flows. These work flows address the uncertainty of vertically and horizontally connected reservoir volumes for productivity at well targets. Model scenarios have been constructed in a 3D visualization environment where optimal integration of a multitude of seismic data volumes, derived attributes, and geological model concepts has been achieved. The roughly polygonally distributed faults are not expected to be sealing; having developed close to sea bed, their origin rules out cataclasis and cataclasis-enhanced cementation. The common gas gradient and absence of measurable depletion during well tests support non-sealing faults and vertical connectivity. However, dynamic fault seal uncertainties related to reservoir heterogeneity and compartmentalization have necessitated risking the relatively simple tank scenario and a more cautious, stepwise approach for the development concept. A significant opportunity can be realized if the gas can be produced profitably using only three templates.

Abstract A consistent design of experiments ( DoE ) based evaluation process was used to assess the magnitude of OOIP uncertainty as well as the relative contributions from uncertainty sources as a function of the historical development of the Jurassic Humma Marrat carbonate reservoir in the Partitioned Neutral Zone ( PNZ ). Within the Marrat interval, three stratigraphic layers, known informally as the A, C, and E zones, produce oil. Porosity and permeability is best developed in the dolomitized lowermost Marrat E interval. Based on limited data, approximately 80-85% of the current oil production is from the E zone and 10-15% from the A zone. The C zone contribution is 5% or less. The uncertainty sources used in the DoE -based evaluation were: structure (time-to-depth conversion and overall interpretation uncertainty), original oil-water contact ( OOWC ), porosity histogram, and oil saturation histogram. All of the uncertainties except structure were evaluated for each of the three stratigraphic zones known to produce oil in the Marrat. High, mid, and low-case values were determined for each of the uncertainty sources listed using well log, core, and analog information available after each well was drilled or as significant new data became available ( e.g. , reprocessed seismic volume). The time period covered by this historical look-back is from 1998 (pre-drill) to 2005. The pre-drill P 50 OOIP estimate for the Humma Marrat was about 900 million reservoir barrels. Following drilling of the initial two wells, the P 50 OOIP estimate was < 400 million reservoir barrels. Subsequent drilling and structure modifications (interpretation and time-to-depth conversion) increased the P 50 OOIP estimate to just over 1500 million reservoir barrels in April 2004. The P 50 OOIP was dropped to 625 million reservoir barrels after Well F was drilled in mid-2005. The OOIP uncertainty range, defined as the P 90 OOIP value minus the P 50 OOIP value, decreased from nearly 700 million barrels in 2004 to about 130 million barrels in mid-2005. Analysis of the DoE -based results show that the significant contributors to OOIP uncertainty changed as additional wells were drilled or existing data was re-processed or reinterpreted. However, the structure and/or OOWC uncertainties were generally the largest, though not necessarily always statistically significant contributors to OOIP uncertainty (based on a 95% confidence level). A normalized uncertainty index ( UI ) derived from the probabilistic OOIP values is used to discuss delineation efficiency and may be useful in delineation well planning.

Abstract An integrated workflow has been developed that successfully characterized a new gas reservoir in the southern Gulf of Mexico, which is utilized to improve reserves estimation, reservoir development, and management planning. The reservoir intervals were contained within a faulted rollover anticline. Based upon development of a sequence stratigraphic framework, the reservoirs were identified as retrogradational shoreface parasequences sitting atop third-order sequence boundaries. AVO and spectral analysis of the seismic volume provided support for this interpretation. A new play concept was developed which incorporated sequence stratigraphy and analysis of 3D seismic attributes for more regional mapping in the area. Recommendations for well stimulation also were made based upon stratigraphic aspects of the rocks.