Abstract The productivity of wells in fractured reservoirs depends, in terms of rate and sustainability, on the heterogeneity and variable connectivity of the open fracture network. Outcrop studies in Cretaceous carbonates from the Catalan Pyrenees illuminate this issue and reveal the degree of uncertainty associated with the interpretation of fracture data from wells and seismic. Three examples are chosen to provide verifiable data, parameters and concepts which can be applied to the workflow of fractured reservoir characterization. We discuss fracture properties and distributions in the subseismic volume, the coupled behaviour between litho-mechanical properties, in situ stress and fracturing, and the permeability properties of fault damage zones. The outcrops also highlight some of the difficulties involved in constructing static reservoir models and evaluating fracture interpretations derived from software-based techniques such as surface curvature.

Abstract The Kayenta Formation, Warner Valley, Utah, shows lateral and vertical clustering of mud-delta propagating-channel sand bodies within a matrix of fine-grained open-lake deposits and further provides opportunity to develop a fluvio-deltaic depositional-process model. Clustering due to nonrandom stream avulsion is well documented for high-accommodation fluvial systems operating in alluvial plains but not well established for lacustrine systems with abundant fluvio-deltaic lobes. Kayenta Formation delta lobes have similar spatial clustering to those observed in fluvial channel belts and possibly extend this clustering concept to fluvio-lacustrine systems. Lithofacies were mapped on three large photo panoramas, and architectural-element analysis was used to identify bounding surfaces of fluvial channel-deltaic lobes. Fluvio-lacustrine delta lobes reflect linear channels that propagate across mud deltas with negligible bifurcation and generate fluvial channel belts incised into lake mudstone. Channels are associated with thin sand sheets or “blow-out wings” that extend multiple channel widths from the channel and cover levee and mud-delta deposits, but delta-front sand beds are absent. The stages of evolution for these propagating channels is preserved in the variation of channel-lobe architecture and reflects mud-delta propagation at the mouth of each channel in the absence of delta-front sand. Sand is outpaced by mud in the ever-lengthening channel, which reduces sand at the channel mouth and diminishes necessity for channel bifurcation. The resulting deposit is thus a frontal mud-delta lobe bisected by a later single sandy channel belt with lateral sand wings. Statistical analysis of these channel belts shows clustering. Clustering of fluvial bodies within shallow lakes is significant in predictive reservoir models because it improves connectivity and localization of delta-lobe reservoirs. The clustering of delta lobes in fluvio-lacustrine systems is theorized to reflect the preferential channel avulsion centered on the axis of the primary channel feeding into the lake and preferential avulsion fairways of feeder channels. Both the segregation of sand and mud though channel lengthening and the clustering are explainable in fully autocyclic terms. The needed allocyclic driver to trigger these fluvio-lacustrine processesis an accommodation rate sufficiently low compared with lake filling rate as to maintain shallow-water conditions across the lake system through multiple generations of channel propagation.

Abstract The Ormen Lange Field is a gas reservoir offshore mid-Norway, developed in a combined structural–stratigraphic–hydrodynamic trap. The lobe-dominated turbidite deposits are mostly of excellent quality, but show a significant deterioration trend towards the fan fringe at its northern margin. Axial parts of the fan contain amalgamated sand-rich deposits, which pass laterally into layered sequences characterized by intercalation of low-permeability heterolithic drapes. Along its 40 km length, the field contains in excess of 400 linked polygonal faults attributed to de-watering of underlying shales. Despite pervasive faulting, reservoir connectivity on a geological timescale is proved by a common pressure gradient in pre-production wells and depletion seen in all later development wells. Recent appraisal drilling of the fan fringe, occupying the crest of the field, encountered only residual gas saturations, despite being located in an area delineated by a seismic direct hydrocarbon indicator. A hydrodynamic aquifer concept is the most plausible explanation for the fluid distribution, in which the gas from the crest of the structure is displaced, leaving behind a northward-thickening prism of residual gas. Dynamic simulation of the fluid-fill evolution over geological time showed the hydrodynamically tilted contact depends on rate of water flow across the aquifer, stratigraphic baffling and faulting, and reservoir quality, i.e. clean sand fraction and effective permeability. Optimal development of this deep-water reservoir depends on understanding the relationship between reservoir quality, connectivity, and the position of the free water level (FWL) in the field. A range of FWL in the north of the field, only weakly constrained by the wells, was empirically established from the hydrodynamically initialized models. This allowed a robust test of the production wells planned to drain the margin of the field. Modelled predictions of reservoir quality and pressures were confirmed by subsequent drilling.

Abstract The Volund Field lies in the Norwegian sector of the North Sea (Quad 24/9). This field produces from a ‘classic’ large-scale sandstone injection complex located in Lower Eocene strata. The sandstone reservoir has been injected into the lower permeability surrounding mudstones of the Sele and Balder formations and Hordaland Group to create an ‘intrusive trap’. The Volund Field consists of a deeper central unit of stacked sandstone sills, surrounded by shallower, steeper-dipping injected sandstone dykes, which make excellent reservoirs with consistently high porosity and permeability. Many of the steeply-dipping injected dykes appear to have excellent connectivity from the water leg through the oil leg and into the gas cap. The complex was identified on seismic data that exhibit a Class 3 amplitude versus offset (AVO) signature on the far-offset stack reflection seismic volume. The seismic data have been used to successfully locate horizontal production wells. Volund seismic geobodies have been extracted and incorporated into the reservoir geomodel to determine the geometry of the injectite features and to populate sands within the injection complex. Volund Field (estimated mean gross resource of 54 mmboe (million barrels oil equivalent)) is producing oil from four horizontal branches (end December 2012), with one water injector well, and has a common oil–water contact and gas–oil contact.

Abstract The Fram discovery, located in the UK Central North Sea, comprises the Paleocene-aged Forties Sandstone Member with an oil rim and primary gas-cap trapped within a four-way dip closure around a pierced salt diapir. The Forties Sandstone Member reservoir at Fram is characterized by very-fine- to fine-grained sandstones interbedded with shales with post-depositional small-scale slumping and sand injection, interpreted to be the product of high-density turbidity currents and debris flows. Deposition was in an overall distal and marginal, basin-floor lobe environment. The Forties reservoir interval is considered to comprise a series of offset-stacked, turbidite lobes characterized by a systematic variation from axial amalgamated sandstone facies to more distal, marginal and thinner-bedded heterolithic sandstone facies, producing an overall sheet-like reservoir architecture. The Forties reservoir at Fram is thinner and poorer when compared with more proximal parts of the Forties submarine fan system, and reservoir quality is strongly controlled by sedimentary facies. The architecture of the reservoir is expected to result in poorer vertical, but greater lateral, stratigraphic continuity when compared with more channelized Forties reservoirs such as the Nelson and Forties fields further to the north. A key step in understanding and characterizing the Fram reservoir was the appraisal drilling in 2009, which included coring, comprehensive wireline logging, formation pressure data acquisition and a drill stem production test. This paper provides an overview of the Fram reservoir geology and demonstrates how integration of data acquired in the 2009 29/3c-8,8z appraisal wells with 3D seismic datasets, existing E&A wells and analogues has helped to improve reservoir characterization and identify the major subsurface uncertainties needing to be addressed during the field-development planning.

Abstract The Merganser Field is located in the East Central Graben of the UK Central North Sea, and consists of a gas/condensate column trapped in a structural attic on the flanks of a fully penetrating salt diapir. A large salt overhang obscures the field and structural definition is challenging owing to poor seismic imaging. Exploration drilling established a column of 1327 ft in Paleocene-aged deep-marine deposits of the Forties, Andrew and Maureen Sandstone members, and revealed significant geological complexity. Depositional styles record the relationship between salt tectonics and sedimentation, with variable reservoir distribution influenced by halo-kinetically induced palaeorelief and accommodation space. Re-mobilization of sediments is observed at multiple scales, and includes centimetre-scale de-watering structures, decimetre-scale sand injectites and kilometre-scale olistoliths. Whilst the hydrocarbon properties are consistent, contact depths are variable. Pressure data indicate compartmentalization across the field, which is likely to be caused by either radial faulting or hydrodynamic effects. Owing to the magnitude of subsurface uncertainty, the Merganser discovery could not sustain the investment required for standalone facilities. The development of the neighbouring Scoter Field provided the requisite local infrastructure to progress Merganser into production. The Field Development Plan (FDP) estimated recovery of 100 bscf gas and 3 mmstb condensate, and focused on delivering a low-cost development solution consisting of two horizontal wells and a subsea tie-back that would be robust against the downside, yet maintain flexibility to optimize in an upside outcome. Pilot holes were drilled to establish top reservoir with the subsequent horizontal well trajectories being re-designed to reflect structural geometry. The reservoir sections would maximize connectivity between fault compartments and stratigraphic units, and positioning was optimized with well-site biostratigraphy. Each reservoir section exceeds 4000 ft and maintains at least 1000 ft vertical stand-off from the gas/water contact. The facilities include a 5 km subsea tie back to the Scoter production manifold, with metering at the Merganser manifold for allocation purposes. Gas and condensate are commingled with Scoter, and transported 11 km to Shearwater for processing. The gas is transported onshore through the Shearwater Elgin Area Line and condensate through the Forties Pipeline System to Kinneil. Field performance to date has exceeded the FDP P50 both in terms of daily rate and cumulative production. Early production rates peaked at 100 mmscf/day of gas and 6000 stb/day of condensate and, to end-2014, Merganser has produced 161 bcf of gas and 10 mmstb condensate. This performance is due to a combination of better than expected connectivity, high reservoir k h , lower draw-down afforded by long horizontal wells and compression at the Shearwater platform. Subsurface uncertainties prior to development were considerable and the ‘appraisal through development’ strategy has demonstrated that success is achievable through meticulous planning and scenario analysis.

The unsaturated zone at Yucca Mountain was investigated as a possible site for the nation's first high-level nuclear waste repository. Scientific investigations included infiltration studies, matrix properties testing, borehole testing and monitoring, underground excavation and testing, and the development of conceptual and numerical models of the hydrologic processes at Yucca Mountain. Infiltration estimates by empirical and geochemical methods range from 0.2 to 1.4 mm/yr and 0.2–6.0 mm/yr, respectively. Infiltration estimates from numerical models range from 4.5 mm/yr to 17.6 mm/yr. Rock matrix properties vary vertically and laterally as the result of depositional processes and subsequent postdepositional alteration. Laboratory tests indicate that the average matrix porosity and hydraulic conductivity values for the main level of the proposed repository (Topopah Spring Tuff middle nonlithophysal zone) are 0.08 and 4.7 × 10 −12 m/s, respectively. In situ fracture hydraulic conductivity values are 3–6 orders of magnitude greater. The permeability of fault zones is approximately an order of magnitude greater than that of the surrounding rock unit. Water samples from the fault zones have tritium concentrations that indicate some component of postnuclear testing. Gas and water vapor movement through the unsaturated zone is driven by changes in barometric pressure, temperature-induced density differences, and wind effects. The subsurface pressure response to surface barometric changes is controlled by the distribution and interconnectedness of fractures, the presence of faults and their ability to conduct gas and vapor, and the moisture content and matrix permeability of the rock units. In situ water potential values are generally less than −0.2 MPa (−2 bar), and the water potential gradients in the Topopah Spring Tuff units are very small. Perched-water zones at Yucca Mountain are associated with the basal vitrophyre of the Topopah Spring Tuff or the Calico Hills bedded tuff. Thermal gradients in the unsaturated zone vary with location, and range from ~2.0 °C to 6.0 °C per 100 m; the variability appears to be associated with topography. Large-scale heater testing identified a heat-pipe signature at ~97 °C, and identified thermally induced and excavation-induced changes in the stress field. Elevated gas-phase CO 2 concentrations and a decrease in the pH of water from the condensation zone also were identified. Conceptual and numerical flow and transport models of Yucca Mountain indicate that infiltration is highly variable, both spatially and temporally. Flow in the unsaturated zone is predominately through fractures in the welded units of the Tiva Canyon and Topopah Spring Tuffs and predominately through the matrix in the Paintbrush Tuff nonwelded units and Calico Hills Formation. Isolated, transient, fast-flow paths, such as faults, do exist but probably carry only a small portion of the total liquid-water flux at Yucca Mountain. The Paintbrush Tuff nonwelded units act as a storage buffer for transient infiltration pulses. Faults may act as flow boundaries and/or fast pathways. Below the proposed repository horizon, low-permeability lithostratigraphic units of the Topopah Spring Tuff and/or the Calico Hills Formation may divert flow laterally to faults that act as conduits to the water table. Advective transport pathways are consistent with flow pathways. Matrix diffusion is the major mechanism for mass transfer between fractures and the matrix and may contribute to retardation of radionuclide transport when fracture flow is dominant. Sorption may retard the movement of radionuclides in the unsaturated zone; however, sorption on mobile colloids may enhance radionuclide transport. Dispersion is not expected to be a major transport mechanism in the unsaturated zone at Yucca Mountain. Natural analogue studies support the concepts that percolating water may be diverted around underground openings and that the percentage of infiltration that becomes seepage decreases as infiltration decreases.