Seismic Characterization of Large-scale Sandstone Intrusions
Published:January 01, 2007
- PDF LinkChapter PDF
Mads Huuse, Andrew Hurst, Joe Cartwright, Noralf Steinsland, 2007. "Seismic Characterization of Large-scale Sandstone Intrusions", Sand Injectites: Implications for Hydrocarbon Exploration and Production, Andrew Hurst, Joseph Cartwright
Download citation file:
Postdepositional remobilization and injection of sand are important processes in deep-water clastic systems. Features resulting from these processes are particularly well documented in the Paleogene of the central and northern North Sea, where large-scale sandstone intrusions significantly affect reservoir geometries and fluid-flow properties of sand and mudstone intervals throughout large areas. Large-scale sandstone intrusions seen in seismic data from the Paleogene of the North Sea can be grouped into three main categories based on their size, morphology, and relation to their parent sand body:
Type 1: Winglike sandstone intrusions are seen as discordant seismic anomalies that emanate from the sides and sometimes from the crests of steep-sided concordant sand bodies, which may be of depositional or intrusive origin. The intrusions may be as much as 50 m (164 ft) thick, and crosscut some 100–250 m (330–820 ft) of compacted mudstone section at angles between 10 and 35°. Winglike intrusions may form regardless of preexisting structures, but commonly exploit polygonal fault systems in the encasing mudstones.
Type 2: Conical sandstone intrusions are seen as conical amplitude anomalies that emanate some 50–300 m (164–1000 ft) upward from distinct apexes located a few meters to more than 1 km (0.6 mi) above the likely parent sand body. The intrusions may be as much as 60 m (196 ft) thick and are discordant to bedding along most of their extent, with dips ranging from 15 to 40°. The nature of the feeder system is conjectural, but may comprise subvertical zones of weakness such as blowout pipes or polygonal fault planes, whereas the intrusions themselves do not appear to be controlled by preexisting fault systems.
Type 3: Crestal intrusion complexes comprise networks of intrusions above more massive parent sand bodies. These intrusions are either too thin or too geometrically complex to be well imaged by seismic data. Despite the small scale of their component intrusions, crestal intrusion complexes may be volumetrically important.
Large-scale sandstone intrusions commonly terminate at unconformities such as base Balder (uppermost Paleocene), top Frigg (lower Eocene), or base Oligocene, where they may have extruded onto the paleo-sea-floor. Because sandstone intrusions are commonly highly porous and permeable, they are important as reservoirs and as efficient plumbing systems in thick mudstone sequences. Because the intrusions occur in unusual stratigraphic positions not predicted by standard sedimentary facies models, they may constitute drilling hazards by hosting shallow gas accumulations or by acting as sinks to dense and overpressured drilling fluids. Predrill prediction of the occurrence of large-scale sandstone intrusions based on seismic data and predictive models is thus vital to successful exploration of deep-water clastic plays.
Until recently, large-scale sandstone intrusions and intrusion complexes were only known from outcrops (e.g., Newsom, 1903; Jenkins, 1930; Smyers and Peterson, 1971; Winslow, 1983; Surlyk, 1987; Parize and Fries, 2003), and their (potential) significance for petroleum systems was only rarely mentioned and commonly ignored ( Jenkins, 1930). Improved coring techniques and advances in three-dimensional (3-D) seismic data quality through the last decade have allowed many examples of large-scale sandstone intrusions to be recognized in the Paleogene of the North Sea Basin, where the intrusions directly impact the morphology and distribution of important reservoir units (Figure 1, Figure 2, Figure 3) ( Jenssen et al., 1993; Newman et al., 1993; Newton and Flanagan, 1993; Timbrell, 1993; Dixon et al., 1995; Lonergan and Cartwright, 1999; MacLeod et al., 1999; Lonergan et al., 2000; Duranti et al., 2002; Purvis et al., 2002; Hurst et al., 2003a, b; Huuse et al., 2003, 2004, 2005a; Duranti and Hurst, 2004; Huuse and Mickelson, 2004). The recent focus on the hydrocarbon-systems impact of large-scale sandstone intrusions has led to the rediscovery that sandstone intrusions may provide important fluid- migration paths through sealing sequences ( Jenkins, 1930; Lonergan et al., 2000; Parnell, 2002; Hurst et al., 2003b; Mazzini et al., 2003; Schwartz et al., 2003; Huuse and Cartwright, 2004; Huuse et al., 2005a).
Three main types of sandstone intrusions have been recognized on seismic data (Figure 4):
Winglike sandstone intrusions: They are seen as crosscutting seismic amplitude anomalies emanating upward from the sides and occasionally from the crests of isolated, steep-sided sand bodies, which may be of depositional or intrusive origin.
Conical sandstone intrusions: They are seen as conical amplitude anomalies emanating upward from discrete apexes, commonly high above the underlying parent sand body.
Crestal intrusion complexes: They make up a crestal intrusion halo above more massive parent sand bodies. These are not always apparent on even good- quality 3-D seismic data, but may be recognized as zones of increased seismic amplitudes and/or heterogeneity correlating with meter-scale intrusions in boreholes.
Although seismic-scale sandstone intrusions have mainly been documented from the North Sea Paleogene and also recently from the Upper Cretaceous ( Jackson, 2007), there is ample unpublished and anecdotal evidence for sandstone intrusions in other deep-water hydrocarbon provinces such as the Cenozoic Atlantic margins offshore northwest Europe, west Africa, and Brazil.
This chapter reviews the seismic characterization of large-scale sandstone intrusions and briefly outlines their significance for hydrocarbon exploration and production. The examples are drawn from the Paleogene of the North Sea Basin, where such features are well known and calibrated by boreholes. The range of intrusive features and their implications for exploration and production are highly pertinent to other deep-water provinces, such as the Atlantic continental margins.
Geological Setting of Large-Scale Sandstone Intrusions in the North Sea
The Cenozoic North Sea Basin is a largely unfaulted epicontinental basin developed above failed Mesozoic rift structures (Figure 1, Figure 2) (Ziegler, 1990). The basin, which maintained relatively deep water along its central axis during most of the Cenozoic, was gradually filled by sediments supplied from the British Isles in the west, the Fennoscandian highlands to the northeast, and, in the late Cenozoic, from the central European landmass to the south (Ziegler, 1990; Huuse, 2002). The North Sea Basin is located immediately to the south of the northwest European Atlantic margin, which underwent continental breakup at the Paleocene–Eocene transition and subsequently went through several episodes of subsidence and inversion during the Cenozoic (Ziegler, 1990; Dore et al., 1999; Davies et al., 2004, sand-rich fans in the upper Paleocene and more isolated, albeit still tens-of-meter-thick, sandstones in the Eocene section (Figure 3) (Den Hartog Jager et al., 1993; Ahmadi et al., 2003; Jones et al., 2003). The sandstones constitute a highly prolific hydrocarbon play in the northern and central North Sea (e.g., Bain, 1993; Jones et al., 2003) and occur in a succession of hemipelagic smectite-rich mudstones that are extremely fine grained (Thyberg et al., 2000). The smectit- ic mudstones are very poorly permeable and, thus, form efficient seals and baffles to fluid migration (e.g., Jones et al., 2003). The uppermost Paleocene–Eocene mudstones are offset by pervasive polygonal fault systems caused by layer-bound contraction and fluid expulsion (Cartwright, 1994; Cartwright and Lonergan, 1996). The sandstone intrusions found in the upper Paleocene–Eocene of the North Sea Basin are the largest scale sandstone intrusions recorded thus far, and it has been argued that their formation is somehow linked to the occurrence of polygonal fault systems (e.g., Lonergan et al., 2000).
Outcrop analogs for large-scale sandstone intrusions comprise localities in East Greenland (Surlyk and Noe-Nygaard, 2001), southern France (Parize and Fries, 2003), and central California (Boehm and Moore, 2002; Schwartz et al., 2003; Thompson et al., 2007). These examples are all characterized by fine- to coarse-grained sandstones encased in very fine-grained mudstones deposited in relatively deep water. Hence, it appears that sandstones encased within deep-water mudstones are particularly susceptible to large-scale remobilization, and it would seem likely that large-scale sandstone intrusions might also be common in other deep-water provinces such as the Cenozoic deep-water successions offshore west Africa and Brazil.
Seismic Imaging of Sandstone Intrusions in the North Sea Paleogene
The dimensions of injected sand bodies encountered both in outcrop and in the subsurface range from millimeter to kilometer scale (Newsom, 1903; Smyers and Peterson, 1971; Dixon et al., 1995; MacLeod et al., 1999; Surlyk and Noe-Nygaard, 2001; Jolly and Lonergan, 2002; Molyneux et al., 2002; Huuse et al., 2004). Because seismic data have limited bandwidth, they are limited in their resolution, and hence, only intrusions or intrusion complexes thicker than 1 m (3.3 ft) or so may be detected by even high-quality 3-D seismic data. The thickness of even large-scale injected sand bodies is commonly around or below seismic resolution (defined as a quarter of the dominant wavelength, λ/4), although commonly above the detection threshold (commonly about λ30), making it very difficult to quantify the volumes of intrusions and of hydrocarbons present within them.
The seismic expression of large-scale injected sand bodies is determined by the interplay between the acoustic properties of the sand and the encasing shales, sand-body geometry and complexity, fundamentals of seismic wave propagation and energy attenuation, and limitations imposed by seismic acquisition and processing techniques. This section provides an overview of specific seismic imaging issues that the authors have found to be relevant to the seismic characterization of large-scale sandstone intrusions in the Paleogene of the North Sea, but the considerations should also be applicable to large-scale sandstone intrusions in deep- water clastic environments. The account given here is aimed at geologists and nonspecialist seismic interpreters; it is by no means a complete treatment of seismic imaging, and interested readers are referred to the large body of published literature on the subject (see Yilmaz, and references therein).
Impedance Contrasts between Sandstones and Mudstones in the North Sea Paleogene
A contrast in acoustic impedance (AI) is required to produce a seismic reflection from an interface. The AI of sandstones and mudstones generally increases with depth, but at different rates (Figure 5), determined by initial porosity and mineralogy and diagenesis, including physical and chemical consolidation. A large number of theoretical and empirical relations exist that relate the velocity or density of sandstones and mudstones to their porosity and/or depth of burial, and the applicability of generalized relations is not always great (e.g., Hilterman, 2001).
For North Sea mudstones, it has been shown that the porosity decrease with increasing burial may be described as a simple exponential decrease with depth (Sclater and Christie, 1980): Ф(z) = Ф0e−cz, where Ф is the porosity, z is the depth (meters), and c is a lithology- dependent compaction factor. Sclater and Christie (1980) used values of Ф0 = 0.63 and c = 0.27 for North Sea mudstones, and Ф0 = 0.49 and c = 0.51 for North Sea sandstones. However, they were considering the entire Mesozoic-Cenozoic succession with one set of parameters that may not be optimal for the shallow (Paleogene) section, which is smectite rich and, thus, likely to have been characterized by higher depositional porosities than average mudstones (cf. Velde, 1996). According to our experience, which includes case studies from the outer Moray Firth, the south Viking Graben, and the north Viking Graben, actual sediment porosities at typical reservoir depths (1.5–2.5 km; 0.9–1.5 mi) are best approximated using Ф0 = 0.8 and c = 0.4 for mudstones and Ф0 = 0.5 and c = 0.2 for sandstones (Figure 5a).
The density of sedimentary rocks depends primarily on the porosity and pore fluids, whereas the velocity also depends on the strength of the grain framework and, to some degree, on fracture density and orientation and on pore-fluid pressure. Simplified relations exist between depth, density, and velocity that may be used to calculate the AI variation with depth for mudstones as well as brine- and oil-saturated sandstones. To calculate the general AI variation with depth, we calculate the bulk density (ρ) as: ρ = Фρfl + (1 − Ф)ρma, where p is the bulk density (g/cm3), Ф is the porosity fraction, pfl is the fluid density (g/cm3), and ρma is the matrix density (g/cm3). We use ρ = 2.72 g/cm3 for mudstones and ρ = 2.65 g/cm3 for sandstones, and ρ = 1.0 g/cm3 for brine and ρ = 0.9 g/cm3 for oil.
The bulk velocity (V ) of a sedimentary rock of a given porosity (Ф) may be calculated using the Wyllie time-average equation: 1/V = (1 − Ф)/Vma + Ф/Vfl, where Vma is the matrix velocity, and Vfl is the fluid velocity. Values for unconsolidated sandstones and mudstones are poorly determined, but we find that Vma = 3 km/s (1.8 mi/s) for both sandstones and mudstones and Vfl = 1.5 km/s (0.9 mi/s) for brine and Vfl = 1.3 km/s (0.8 mi/s) for oil results in reasonably good correspondence with the observed bulk velocities, when the modified parameters are used to calculate the porosity as a function of depth using the relation given above. The generalized velocity and density variations of smectite-rich mudstones with depth are plotted in Figure 5b, and the AI variation with depth is plotted in Figure 5c.
The AI contrast between the target sandstones and the encasing mudstones determines the reflection coefficient, which, in turn, determines the seismic reflection response. For vertical incidence seismic reflections, the reflection coefficient (RC) is calculated as RC = (Alsand AImud)/(AIsand + AImud). In an ideal situation, the reflection coefficient varies predictably with depth, and it can be shown that at certain depths, all other things being equal, sandstones may have no AI contrast with the encasing mudstones (Figure 5d). In reality, other factors, most notably cementation, complicate the picture, and at particular depth intervals, a prospective sandstone may show up as bright positive, bright negative, or dim, depending on porosity variations caused by cementation differences in the sandstone and mineralogy-cementation variations of the mudstones (Figure 6).
At depths of 1.5–2.5 km (0.9–1.5 mi), hydrocarbon- bearing reservoir sandstones in the North Sea Paleogene commonly have poor AI contrasts with the encasing mudstone, leading to poor definition of the top reservoir (e.g., Lonergan and Cartwright, 1999; MacLeod et al., 1999). In many cases, these reservoirs were intersected by chance when drilling for deeper targets, such as tilted Mesozoic fault blocks; and the top reservoir horizon has typically been estimated by assuming it was smooth and parallel to an overlying horizon (Figure 7) (MacLeod et al., 1999; Duranti et al., 2002; Jones et al.,2003). For remobilized reservoirs, this approach results in a poor match between predicted and actual reservoir occurrence, unreliable net-to-gross (N/G) estimates, and nonoptimal placement of production wells (MacLeod et al., 1999; Duranti et al., 2002; Huuse et al., 2003; Jones et al., 2003).
Reservoir imaging may be drastically improved by converted (PS) wave seismic imaging, which involves conducting an Ocean Bottom Cable (OBC) survey to record the shear-wave component of the reflected seismic signal (e.g., MacLeod et al., 1999; Stewart et al., 2002). Converted-wave seismic data commonly highlight sandstones as bright hard events because of the relatively high shear-wave velocity of sandstones relative to mudstones at similar burial depth (Figure 8) (MacLeod et al., 1999). PS data facilitate the mapping of bulk top reservoir, i.e., the top of the massive reservoir sandstone and major sandstone intrusions emanating from it ( Jones et al., 2003), but PS data generally do not provide sufficient resolution to allow mapping of small-scale crestal intrusions. In some cases, most notably for the Alba field, PS imaging has completely changed reservoir models of producing fields, leading to markedly enhanced productivity (e.g., MacLeod et al., 1999; Duranti et al., 2002; Jones et al., 2003). Once the gross distribution of sandstone and mudstone is known, more detailed analyses may focus on the porosity variations of the reservoir sandstones.
Many remobilized sandstones consist chiefly of an ambiguous massive sandstone facies, which may be de- positional, remobilized, or injected in origin (Duranti and Hurst, 2004). In some cases, remobilized and injected sandstones have slightly tighter grain packing and, thus, higher AI than nonremobilized sandstones, and this variation may be detected in sonic and density logs (Duranti et al., 2002). AI variation in a sand body resulting from postdepositional remobilization may be detected inverting the seismic data for AI (Figure 9). The AI data may then be used to optimize reservoir models and production drilling by delineating relatively porous and relatively tight parts of the reservoir that may be significant for, e.g., water coning during production of reservoirs with thin oil columns. Moreover, if it can be established that remobilized sandstones are associated with a relatively high AI response, the AI data may assist in assessing the degree of remobilization and injection of a given reservoir. However, experience shows that such relations need to be established for each individual reservoir, and caution should be taken to assume a general association of high AI response and injectites because variations in cementation may be more important than grain packing for many sandstone intervals.
Seismic resolution is largely a function of spatial sampling density, frequency bandwidth, migration aperture, and noise. Vertical resolution is defined differently by different authors; here, we choose the commonly used definition of resolution as a quarter of the dominant wavelength (λ/4), which is the thickness that results in maximum tuning between reflections from interfaces with opposite reflection coefficients (e.g., Sheriff, 1985). The dominant wavelength is defined as λ = V/f, where V is the interval velocity, and f is the dominant frequency, i.e., the center-frequency of the amplitude-frequency spectrum of the data.
The dominant frequency of the seismic data at reservoir depth is typically 25–50 Hz for surface-recorded P-wave seismic data depending on acquisition and processing parameters and on geological conditions (e.g., presence of gas, abnormally high AI bodies, etc.). Typical frequencies for PS data may range from 15 to 20 Hz. At typical sandstone porosities (˜20–40%), P-wave velocities may range from 2 to 4 km/s (1.2 to 2.5 mi/s), whereas S-wave velocities may be of the order of 1–2km/s (0.6–1.2 mi/s). The resolution (λ/4) of both data modes is thus within the range of 10–40 m (33–131 ft), with 20–30 m (66–100 ft) being typical of the North Sea Paleogene in 3-D seismic data (Figure 10). This yields a threshold for detection, defined as λ/30, of about 5–1 m (1.6–3.3 ft).
The lateral resolution of migrated 3-D seismic data is dependent on the dominant wavelength and the inline and cross-line spacing. For most practical purposes, the resolution is of the order of the dominant wavelength, i.e., about 80–100 m (262–330 ft), although experience shows that features significantly smaller than this may be detected. Some (especially vintage) 3-D seismic surveys acquired with inline spacing greater than 50 m (164 ft) are subject to spatial aliasing, and the lateral resolution of these surveys is thus somewhat coarser than the dominant wavelength.
The imaging limitations of seismic data can be illustrated by conducting forward modeling of sandstone intrusions mapped at outcrop (Figure 11). Even the largest scale outcrop examples currently known are slightly smaller in scale than the largest examples reported from the subsurface, but some remarkable and useful examples of large-scale sandstone intrusions occur in various deep-water successions, spanning Cretaceous- Miocene mudstones onshore central California (Thompson et al., 1999, 2007; Schwartz et al., 2003). For the purpose of illustrating the limitations of seismic imaging of sandstone intrusions, we populated an outcrop section from the Tumey Hills (California) (Figure 11) with subsurface acoustic properties known from cores and well logs of the Alba and Chestnut fields located in the Eocene of the outer Moray Firth (United Kingdom North Sea). The resulting impedance section (Figure 12a) was then convolved with zero-phase Ricker wavelets of varying frequencies (Figure 12b) and subject to diffraction modeling migrated using actual velocities (Figure 12c). The results highlight that realistic seismic frequencies only provide a coarse image of the actual geometries, and that frequencies required are two to four times higher than those typically available in the North Sea Paleogene.
With the steadily increasing evolution of 3-D seismic acquisition techniques, it has become possible to acquire surface-towed 3-D seismic data at much higher spatial resolution than previously feasible. Spatial aliasing in some recent 3-D seismic surveys is thus mainly a matter of inherent limitations in frequency bandwidth, leading to almost twice the resolution of vintage 3-D data in some cases (e.g., McHugo et al., 2003; Long and Buchan, 2004). As illustrated by the modeling example (Figure 12), such recent improvements in seismic bandwidth should prove extremely useful when dealing with injected sandstone reservoirs, especially when combined with modern, prestack migration algorithms.
Imaging Improvements by Advanced Migration Algorithms
Since the advent of digital data recording, the processing of seismic data has been continuously improving, involving careful model building, fine tuning of (migration) algorithms, and innovation of alternative processing flows to achieve optimal subsurface illumination. Recently, it has been demonstrated that vast improvements in imaging quality of potential sandstone intrusions may be achieved by the reprocessing of vintage seismic data, considering the anomalous velocities and steeply dipping geometries associated with sandstone intrusions and anisotropy in the encasing mudstones. The remedies for these problems include building detailed velocity models, performing prestack time or depth migration, and including the effects of seismic anisotropy (cf. Mikhailov et al., 2001; Luchford, 2002; Løseth et al., 2003).
An example of the effects of reprocessing 3-D seismic data using prestack time migration is given in Figure 13, which shows imaging improvements of suspected sandstone intrusions along the flanks of an oil discovery, and complete reevaluation of the geological and reservoir models for the discovery. Careful processing also resulted in better amplitude preservation, allowing the volumetric estimation of subseismic-scale intrusions over the crest of the central sand accumulation. However, regardless of sampling density and accuracy of the seismic migration, the seismic data are still limited in their ability to image steeply dipping features, such as sandstone dikes. One should thus not expect to image sandstone dikes inclined more than 45–60°, depending on their thickness and rugosity relative to the seismic resolution.
Visualization of Large-scale Intrusions
The availability of high-powered computing facilities and a variety of 3-D interpretation and visualization packages has facilitated the interpretation and visualization of a variety of geological structures, not the least of which are sandstone intrusions (Gras and Cartwright, 2002; Jones et al., 2003; Huuse and Mickelson, 2004; Huuse et al., 2004, 2005a). Whether particular sandstone intrusion complexes lend themselves better to volume or line-based interpretation is a function of the imaging issues discussed above and the intricacies of the intrusions themselves (Huuse et al., 2004). In many cases, it is advisable to adopt a dual approach, where initial screening of the data is best performed in the volume visualization environment, whereas some standard line-based horizon interpretation is commonly required before the actual volume-based interpretation is conducted using voxel tracking and growing. To be successful, volume interpretation requires that the target has a well-defined seismic character, and preferably, that some restrictions are imposed on the volume growth algorithm, either by sculpted data volumes or by additional constraints on the interpretation, such as dual attribute picking of, e.g., amplitude and impedance. In some cases, it may be useful to interpret the intrusions using both 3-D and two-dimensional (2-D) approaches because the former allows intricate structures to be mapped in an almost objective manner, whereas the latter is better suited for targets or data with poor amplitude contrast and commonly yield results better suited for comparison with structural trends and polygonal faults in the encasing mudstones. The examples provided in the remainder of this chapter originate from a mixture of line- and volume-based approaches and, thus, highlight the variable expressions of large-scale sandstone intrusions in both seismic data and the derived interpretations.
Seismic Examples Of Large-Scale Sandstone Intrusions
Large-scale sandstone intrusions seen in seismic data from the North Sea may be divided into three main categories, based on their seismic characteristics and relation to (inferred) parent sands, including winglike, conical, and small-scale crestal sandstone intrusions (Figure 14). The three types of intrusions are described in the following, illustrated by examples from the Paleogene of the North Sea.
Winglike Sandstone Intrusions
The term “winglike intrusions” was first used about sandstone intrusions (mainly sills) (Figure 13Figure 14Figure 15Figure 16Figure 17) emanating sideways from Jurassic gully-fill sandstones in the Hareelv Formation of Eastern Greenland (Surlyk, 1987) (Figure 10). Sandstone sills interpreted as caused by lateral injection of sand during gully sand deposition are also found in the Cretaceous blue marls of southern France (Parize and Fries, 2003). Although in both cases, thin dikes are relatively abundant, the thickest intrusions mainly comprise sills that generally have maximum inclinations with respect to bedding of a few degrees and up- and downstepping elements of limited vertical extent (Surlyk, 1987; Surlyk and Noe-Nygaard, 2001; Parize and Fries, 2003). This type of near-concordant winglike intrusion adjacent to turbidite channels has not yet been documented from the subsurface, where the reported wings are markedly discordant along most of their extent.
Winglike sandstone intrusions in the subsurface were first reported from the flanks of the Alba field, located in the Eocene of the outer Moray Firth (Figure 7) (Lonergan and Cartwright, 1999; MacLeod et al., 1999). These and similar features observed elsewhere are seen as discordant reflections inclined a few tens of degrees with respect to bedding (see Figure 13Figure 14Figure 15Figure 16Figure 17). The winglike intrusions seen in the North Sea Paleogene generally outline tens-of-meter-thick, steep-sided, and remobilized sand bodies (Dixon et al., 1995; Lonergan and Cartwright, 1999; Duranti et al., 2002; Hurst et al., 2003a, b; Huuse et al., 2003, 2004). The intrusions range from 10 to 40 m (33 to 131 ft) thickness, 50–250 m (164–820 ft) vertical extent and may be more or less continuous for 1–5 km (0.6–3.1 mi) or more along the edge of the parent sand, depending on the extent of the parent sand body and subsequent modification by polygonal faulting. The lateral distance from the parent sand body may be more than 1 km (0.6 mi) when including the most distal tip of concordant wing segments (Figure 14).
The winglike reflections emanating from the edge of the main Alba sand body have been calibrated by pilot and production wells proving the existence of tens-of-meter-thick inclined sandstone intrusions of excellent reservoir properties, resulting in some of the best producers in the Alba oil field (MacLeod et al., 1999). Recently, several borehole penetrations of winglike anomalies have been reported (Cole et al., 2000; Bergslien, 2002; Huuse et al., 2004; de Boer et al., 2007), all calibrating the anomalies to tens-of-meter-thick massive sandstones. Borehole calibrations typically comprise only a limited suite of petrophysical logs because the intrusions commonly are encountered at anomalous levels, and presently, there is a paucity of core calibrations of winglike intrusions (except de Boer et al., 2007).
The present-day inclinations of the wings are typically of the order of 10–35°. Depending on the compaction state at the time of intrusion, the original intrusion angle would have been somewhat steeper, perhaps as much as 60°, and it would seem most appropriate to classify the winglike intrusions as low- angle dikes instead of transgressive sills. In many cases, the wings can be seen to tip out or become concordant at unconformities in the overburden (Figure 8, 13–15) (Alba: Lonergan and Cartwright, 1999; Grieg, Gamma: Huuse et al., 2004). This could suggest that they terminated by extrusion onto the paleo-sea-floor (Huuse et al., 2004) or intruded concordantly at shallow burial depths (Lonergan et al., 2000; Jolly and Lonergan, 2002). It has been observed that concordant wing tips only protrude away from the parent sand body and wing, and this has been used to invoke extrusion to the paleoseabed (Huuse et al., 2004). However, propagation away from the feeder dike would also be expected from an intrusion obliquely abutting a rheological interface (Pollard, 1973), and the propagation direction thus cannot be used as direct evidence for extrusion versus sill termination of the wings. Hence, the problem of seabed versus subsurface termination of the wings and other large-scale intrusions awaits outcrop analog studies and/or good core calibration of the wing tips.
The origin of some of the steep-sided parent sand bodies from which the winglike intrusions emanate is also enigmatic because they display uncommonly steep sides and commonly lack confinement by levees or ero- sional scours (Figure 13Figure 14Figure 15Figure 16) (Huuse et al., 2004). Assessing the mode of emplacement of the parent sand body is hindered by the commonly extensive remobilization features (Newman et al., 1993; Dixon et al., 1995; Lonergan and Cartwright, 1999; Lonergan et al., 2000; Huuse et al., 2004, 2005 a). Depositional models commonly invoke confined deposition in erosional scours (Newton and Flanagan, 1993; Lonergan and Cartwright, 1999), in ponded basin-floor settings (e.g., Newman et al., 1993; Purvis et al., 2002), in giant pockmarks (Cole et al., 2000), or deposition by freezing of sandy debris flows and turbidites ( Jennette et al., 2000; Bergslien, 2002). In many cases, none of these models adequately explain the occurrence of tens of meters of clean sands without any apparent confinement (Huuse et al., 2004). Hence, because some of the parent sand body and wing geometries bear distinct resemblance with saucer-shaped igneous laccoliths (Cosgrove and Hillier, 2000; Hansen et al., 2004), it might be speculated that some winglike intrusions could have been injected from entirely intrusive sand bodies (laccoliths) sourced from deeper Paleogene sands (see also de Boer et al., 2007).
In some cases, a completely injected origin of the wings and their basal parent sand bodies is contradicted by observations of depositional structures made on cores recovered from some of the steep-sided sand bodies that would seem to preclude an intrusive origin for the entire parent sand body (e.g., Bergslien, 2002; Duranti et al., 2002). However, field studies have shown that laminae may occur in completely intrusive sand bodies (Peterson, 1968; Injected Sands Group, unpublished data). Core-scale structures used to infer a depositional origin of thick sandstones should thus occur in a context and be of a sufficiently unequivocal nature for them to be identified as definite deposi- tional features (see Duranti and Hurst, 2004, for an indepth discussion of sand injectite facies recognized in cores). In summary, depending on context, the concordant sandstone at the base of the winglike anomalies may be of depositional or intrusive origin. The latter would effectively constitute a sandstone laccolith, and the total sandstone volume in such structures (e.g., Figure 14) can be of the order of 0.5 km3 (0.12 mi3).
Conical Sandstone Intrusions
Kilometer-scale conical sandstone intrusions (Figure 17Figure 18Figure 19) were recently discovered in the lower Eocene of the outer Moray Firth, where they are seen as discordant, commonly high-amplitude, V-shaped reflections at typical seismic display scales (Figure 4, Figure 18a) (Lonergan et al., 2000; Gras and Cartwright, 2002; Moly- neux et al., 2002; Huuse et al., 2005a). The reflections are circular to polygonal in plan view, forming inverse conical to pyramidal structures in 3-D. Similar intrusions are abundant in the lower Eocene of the south and north Viking grabens (Figure 18, Figure 19) (Molyneux, 2001; Huuse et al., 2003, 2004; Løseth et al., 2003; Huuse and Mickelson, 2004), and in the Faeroe-Shetland Basin (Huuse et al., 2001; Shoulders and Cartwright, 2004). Several well calibrations of these conical amplitude anomalies exist, tying the anomalous reflections to tens- of-meter-thick tabular sandstones that may be partly cemented (e.g., Molyneux et al., 2002; Løseth et al., 2003; Huuse and Mickelson, 2004; Huuse et al., 2004, 2005a). The well calibrations and the discordant relations with reflections in the encasing mudstones are consistent with an interpretation as large-scale conical sandstone intrusions.
Conical intrusions display a similar or slightly steeper range of dips than winglike intrusions, but they are distinct by their downward termination in a well- defined point or occasionally linear apex, leaving little or no room for a concordant sand body at their base. The upward divergence from the distinct apex suggests that the intrusions were emplaced by an upward intrusion from a discrete point or occasionally linear feeder system. In the outer Moray Firth and in the Tampen Spur region of the north Viking Graben, the distance to the nearest underlying sand is of the order of 50200 m (164–660 ft) (e.g., Huuse and Mickelson, 2004; Huuse et al., 2005a), whereas conical intrusions in the south Viking Graben appear to be located several hundreds of meters above the nearest possible underlying parent sand body (Huuse et al., 2004).
The true geometry of conical intrusions is not known because of the limitations of seismic resolution and the lack of well-exposed outcrop analogs of similar-scale intrusions. Opacity rendering (Figure 18c) and voxel picking of conical intrusions generally yield complex intrusion geometries that are highly fragmented although still consistent with an overall conical shape (Figure 18b). Outcrop modeling of moderately complex sandstone intrusion geometries suggests that such seismic irregularities may originate from true geometric variability of the intruded sandstones, although it is clear that the true sandstone geometries cannot be resolved by the combination of frequency and velocity characteristic of standard 3-D seismic data available for the Paleogene of the North Sea (Figure 12).
Conical intrusions range from 0.5 to 2 km (0.3 to 1.2 mi) diameter (typically ˜1 km [˜0.6 mi]) and 50300 m (164–1000 ft) height. Their calibrated thicknesses are of the order of 10–50 m (33–164 ft) or more. Bulk volumes of conical intrusions may be in the range of 106–108 m3 (0.035–3.5 x 109 ft3), and the largest scale intrusions may thus provide significant reservoir volumes (Huuse and Mickelson, 2004; Huuse et al., 2005a), whereas small- or medium-scale conical intrusions may provide efficient fluid-migration paths through tens to hundreds of meters of mudstone section (Hurst et al., 2003b; Løseth et al., 2003; Shoulders and Cartwright, 2004; Huuse et al., 2005a).
Because conical intrusions may be located hundreds of meters above their parent sand body, they do not fit into depositional models and may occur unexpectedly when drilling for deeper reservoirs. Unanticipated penetration of a large-volume conical intrusion while drilling through smectitic mudstones could result in a variety of drilling problems, including circulation loss (if porous and normally pressured), blowout (if overpressured and gas or oil charged), or decreased penetration rate (if completely cemented). Circulation losses may be particularly dramatic when encountering large- scale conical intrusions in the lower Eocene because of large mud weights commonly used for drilling the polygonally faulted overlying and encasing mudstones.
Crestal Intrusion Complexes
Small-scale crestal intrusions are frequently encountered in borehole cores up to a few hundred meters above massive sandstone accumulations in the North Sea Paleogene (Newman et al., 1993; Dixon et al., 1995; Lonergan et al., 2000; Bergslien, 2002; Duranti et al., 2002; Purvis et al., 2002; Duranti and Hurst, 2004). The intrusions making up a crestal intrusion complex (Figure 13, Figure 20) are typically of centimeter to meter scale, and a wide variety of intrusion styles have been observed in core, ranging from planar dikes and sills to pervasively intruded mudstone clast breccias (e.g., Duranti et al., 2002). An outcrop analog of this type of intrusions may be found in the Pannoche Hills of central California, where planar intrusions a few meters thick crosscut several hundred meters of mudstones (Smyers and Peterson, 1971; Schwartz et al., 2003; Huuse et al., 2004; their figure 4b).
Because of their small thickness and intricate geometries, crestal intrusions are not readily interpreted using seismic data. They are, nevertheless, important because they constitute an important component of the reservoir volume in many oil discoveries and may provide vertical connectivity between stratigraphically separate reservoirs (Dixon et al., 1995; Lonergan et al., 2000; Guargena et al., 2002).
In a recent study, we attempted a volume-based interpretation of a completely injected crestal intrusion complex overlying an approximately 100-m (330-ft)- thick Balder Sandstone mound. It was decided to define the reservoir geometry using voxel picking of reprocessed (prestack depth-migrated, time-converted) 3-D seismic data (Figure 13bFigure 20). This approach was chosen because intrusion geometries captured by 2-D surface- based interpretation rarely captures the most realistic geometries of sandstone intrusions (cf. Figure 18a,b) (Huuse et al., 2004, 2005a). Limited core calibration from the discovery well (Conoco, Production License 039, well 24/9-3) suggests that the reservoir sandstones are all of intrusive origin (D. Duranti, 2002, personal communication) and similar in character to those of the underperforming Leadon field reservoir (Templeton et al., 2002). Root mean square–amplitude extractions of the seismic data suggest that the crestal intrusion fringe penetrated by the discovery well correlates with high amplitudes on the reprocessed seismic data. To visualize the picked voxels, the top and base of the picks at each map location were converted to the enveloping surfaces and color coded according to their likely fluid content (Figure 20). The result highlights the presence of thick, winglike intrusions along the flanks of the discovery and a fringe of bright voxels over the highest parts of the Balder mound (Figure 20). The absence of bright voxels over the crests of the lower parts of the mound to the southeast could be caused by the absence of intrusions, but is probably more likely caused by the absence of hydrocarbons in this location.
Some Remarks on the Origin of Large-Scale Sandstone Intrusions
Large-scale Sandstone Intrusions and Polygonal Faults
The largest scale sandstone intrusions, i.e., the winglike and the conical sandstone intrusions, have so far only been found in polygonally faulted successions of smectite-rich claystones (Lonergan and Cartwright, 1999; Lonergan et al., 2000; Molyneux et al., 2002; Hurst et al., 2003a, b; Huuse and Mickelson, 2004; Huuse et al., 2004, 2005 a; Shoulders and Cartwright, 2004). The plan geometries of polygonal faults and large-scale sandstone intrusions are similar (e.g., Lonergan et al., 2000). Based on this argument and on occasional coincidences between polygonal faults and intrusions in cross section, it has been suggested that they may be genetically related (e.g., Lonergan and Cartwright, 1999; Lonergan et al., 2000; Molyneux et al., 2002; Hurst et al., 2003b). However, the large-scale sandstone intrusions are generally inclined by some 10–35°, whereas polygonal faults in the same tiers generally range from 20 to 60° dip (cf. Lonergan and Cartwright, 1999; Shoulders, 2005). Both dip ranges are modified by compaction of the encasing mudstones and may have been closer to 30–50 and 35–75°, respectively. Moreover, it is exceedingly uncommon that constituent faults defining polygonal fault cells show consistent inward dip and downward termination in an apex, whereas this is a defining characteristic of the abundant conical intrusions (Huuse and Mickelson, 2004; Huuse et al., 2004, 2005a). Finally, it should be noted that relations between intrusions and polygonal faults in cross section comprise coincidences as well as crosscutting relations (Huuse et al., 2001), which would be expected by two coexisting and pervasively developed, but largely independent, discordant systems. Hence, we suggest that the co-occurrence of large-scale sandstone intrusions and polygonal faults in the outer Moray Firth, the south Viking Graben, the north Viking Graben, and the Faroe-Shetland Basin is primarily caused by a common dependence on the rheology and stress state of the smectite-rich host mudstones, instead of a general genetic link between the two phenomena. However, even if there is no general dependence of sand injection upon polygonal faulting, faults that are active during sand injection are probably likely to be at least partly exploited by sand intrusions. Intrusion along polygonal faults may be more common for winglike intrusions that are more closely associated with their encasing mudstones than for conical intrusions, which have been intruded from a variety of depths below the interval in which they are emplaced.
Steep-sided Mounds and Winglike Intrusions
The consistent occurrence of the winglike intrusions along the edges of massive sandstones (Figure 13, Figure 14, Figure 15, Figure 16) has previously been attributed to polygonal faulting (Lonergan and Cartwright, 1999; Lonergan et al., 2000). However, the marginal position could also be explained by forced folding of the overburden caused by differential compaction across a depositional sandstone body (the laccolith model suggested for the Alba field wings by Cosgrove and Hillier, 2000). A more unconventional explanation could be that the basal parent sandstone is also intrusive, and that the winglike anomalies emanate from a true sandstone sill (cf. de Boer et al., 2007). A completely injected origin is contradicted by depositional structures reported from cores of the remobilized Alba field in the outer Moray Firth (Duranti et al., 2002; Duranti and Hurst, 2004). However, apart from the Alba field, such structures are relatively uncommon, and reports of laminae and flow structures in intrusive sandstones (Peterson, 1968; Injected Sands Group, unpublished data) might suggest that some inferred depositional sandstones could, in fact, be of intrusive origin. In addition, it may be speculated that some normally pressured depositional sand bodies could have been inflated by sand intruded at their base from a deeper overpressured sand source, leading to remobilization of the in-situ sand body. Such a scenario could explain the depositional and remobilizational characteristics of several North Sea reservoirs that we have examined.
Because of the wide range of possible variations and a lack of sufficient hard data, the present discussion cannot provide a general conclusion regarding the proportions of depositional versus injected sand bodies found at the bases of winglike intrusions. It should merely be noted that there appear to be cases where the bulk of the concordant basal sand body is most likely of depositional origin (e.g., Alba; Duranti and Hurst, 2004), and there are cases where the entire body is likely to be of injected origin (Figure 13, Figure 14, Figure 15) (de Boer et al., 2007). Care should thus be taken to interpret the sand body origin based on the available data for each case in question.
Driving Force and Triggering of Large-scale Sand Intrusion
The mechanics of large-scale sand intrusion are still poorly known because factors such as the transport mode of the sand and fluid mixture and the depth of parent sand body and intrusion emplacement depths are still subject to debate in most of the cases (cf. Lonergan et al., 2000; Jolly and Lonergan, 2002; Huuse and Mickelson, 2004; Huuse et al., 2004; 2005a, b; Duranti and Mazzini, 2005; Jackson, 2007). However, based on regional compilations and individual case studies, some general remarks can be made regarding the conditions enabling the upward intrusion of millions of cubic meters of sand over large areas in the Paleocene and Eocene of the North Sea Basin. When discussing the mechanics and timing of sand injection events, it is important to consider both background priming conditions facilitating the remobilization of sand bodies and the actual events that trigger the remobilization and injection. The main ingredients required for the generation of large-scale sandstone intrusions on a regional scale include
the widespread occurrence of unconsolidated sand bodies encased in sealing mudstones
the regional overpressure caused by one or more mechanisms, such as disequilibrium compaction, lateral pressure transfer, and fluid (hydrocarbon) buoyancy
a triggering event such as a large magnitude earthquake, a bolide impact, or a widespread faulting episode
A relatively slow buildup of overpressure may be caused by a variety of factors, such as disequilibrium compaction, and lateral transfer (e.g., Osborne and Swarbrick, 1997). Because of the extremely finegrained and poorly permeable nature of the smectite-rich Eocene mudstones in the North Sea, disequilibrium compaction could be an effective means of regional overpressure generation. Lateral transfer is widely responsible for overpressure generation in the shallow section of the Gulf of Mexico (e.g., Ostermeier et al., 2001) but would be less likely to affect the basin central parts of the North Sea Eocene where the giant intrusions occur in abundance. Lateral transfer should be most effective toward the margin of the North Sea because of the basinward inclination of the extensive Paleocene aquifers. The occurrence of large-scale intrusions along the axis of the central and northern North Sea has led to the suggestion that overpressures could have been provided by hydrocarbon migration into shallow unconsolidated reservoirs (Jolly and Lonergan, 2002; Mazzini et al., 2003). However, hydrocarbon-related seep structures have only been identified in one core from the North Sea Eocene (Mazzini et al., 2003; Duranti and Mazzini, 2005), and it is thus uncertain whether hydrocarbon migration and/or buoyancy is applicable as a mechanism for causing regional overpressures in the Eocene of the central and northern North Sea. Based on these considerations, it seems that disequilibrium compaction would be the most likely cause of regionally developed overpressures, which may have been exacerbated locally by fluid buoyancy and/or lateral transfer of fluids.
Many studies assume that sandstone intrusions generally tip out at or close to the paleoseabed (e.g., Jolly and Lonergan, 2002; Huuse et al., 2004). Following this assumption, we propose that the large-scale sandstone intrusions in the North Sea Basin record at least three major phases of sand remobilization and injection:
earliest Eocene (top Frigg) injection of winglike intrusions and crestal intrusion complexes in the south Viking Graben (Figure 13, Figure 14, Figure 15, Figure 16) (Dixon et al., 1995; Huuse et al., 2004)
middle Eocene (blue marker) injection of conical sandstone intrusions throughout the axis of the central and northern North Sea (Figure 8, Figure 18, Figure 19) (Molyneux et al., 2002; Løseth et al., 2003; Huuse and Mickelson, 2004; Huuse et al., 2004, 2005a)
A transient buildup of overpressure could occur because of liquefaction of unconsolidated and overpressured sands by earthquake or impact shaking (Alvarez et al., 1998; Obermeier et al., 2002; Huuse and Mickelson, 2004). Most engineering studies suggest that earthquakes do not generally cause liquefaction below a few tens of meters burial (e.g., Obermeier et al., 2002). However, engineering studies are limited in their observational database by only including the uppermost few tens of meters of the subsurface and only include terrestrial settings. Recent studies of an earthquake site in Canada suggest that sands may have been liquefied to a depth of 100 m (330 ft) (Benjumea et al., 2003), and paleoliquefaction structures in Utah suggest minimum depths of liquefaction in excess of 100 m (330 ft) (Alvarez et al., 1998; Netoff, 2002; Huuse et al., 2005b). These examples are from continental settings where preexisting overpressures are unlikely to have been developed. In the presence of substantial, preexisting overpressures, it is thus conceivable that an earthquake could cause liquefaction at depths significantly greater than 100 m (330 ft), in particular in settings where sand bodies are abundant and the overburden is highly porous and water saturated. Because deep-marine (slope) sediments are prone to disequilibrium compaction and commonly overlie hydrocarbon-prone basins, they may be particularly prone to remobilization and injection of sand, as also suggested by the numerous examples of sandstone intrusions in marine slope sediments. Because there are no known impact craters of the appropriate age in the vicinity of the North Sea Basin, we think it most likely that the intrusion events would have been triggered by large-magnitude earthquakes in relation to the North Atlantic tectonic compression events (cf. Muir Wood, 1989; Dore et al., 1999; Huuse and Mickelson, 2004; Huuse et al., 2004, 2005a).
Implications for Exploration and Production
The implication of large-scale sandstone intrusions for hydrocarbon systems has been known for over a century and is well documented from outcrops in California (Newsom, 1903; Jenkins, 1930). Large-scale sandstone intrusions in the subsurface were first discovered in connection with the prolific Paleogene sandstone play of the northern North Sea (Newman et al., 1993; Timbrell, 1993; Dixon et al., 1995; Lonergan and Cartwright, 1999; Duranti et al., 2002; Hurst et al., 2003a, b; Huuse et al., 2003, 2004, 2005a; Huuse and Mickelson, 2004). The occurrence of thick, highly porous, and permeable sandstone bodies crosscutting hundreds of meters of poorly permeable mudstones has obvious implications for reservoir geometry and connectivity and for petroleum migration (Figure 4) (Dixon et al., 1995; Lonergan et al., 2000; Duranti et al., 2002; Hurst et al., 2003a, b; Huuse et al., 2003; Huuse and Cartwright, 2004):
deformation of top reservoir and displacement of reservoir rock into crestal intrusion fringes that may be mistaken for ratty turbidites hundreds of meters above underlying massive sandstones
winglike intrusions that may contain significant upside potential in a lateral or downflank position
formation of intra- and interreservoir connectivity by breach of intra- and interreservoir mudstone seals
Less obvious effects of sand remobilization and injection on reservoir quality vary from a slight decrease in porosity and permeability caused by in-situ remobilization (Duranti et al., 2002) to pervasive lithology mixing caused by brecciation of the host mudstones and forceful intrusion of sand (Duranti et al., 2002; Templeton et al., 2002).
Large-scale sandstone intrusions are still considered unconventional, if considered at all, by most of the exploration community and are thus commonly encountered at surprising levels in the overburden when drilling for deeper, more obvious and conventional prospects (Hurst et al., 2005; Huuse et al., 2005a). Because their formation is poorly understood and their widespread occurrence is only starting to be realized, discordant amplitude anomalies are commonly considered artifacts and ignored or avoided by explorationists. However, apart from potential hydrocarbon targets, large-scale sandstone intrusions may also constitute significant drilling hazards (Huuse et al., 2005a). Because the intrusions contain large pore volumes and occur within polygonally faulted mudstone sequences that are commonly drilled with overbalanced mud weights, they may cause significant loss of drilling fluid when intersected by boreholes.
Large-scale sandstone intrusions are thus significant to the exploration and production of hydrocarbons for several reasons:
Sandstone intrusions may constitute volumetrically significant reservoirs (intrusive and extrusive traps).
They impact reservoir quality, distribution, and connectivity.
They provide vertical and lateral connectivity across long distances (hundreds of meters to kilometers).
They constitute hazards to drilling operations.
To maximize the exploration and production potential of hydrocarbon systems affected by large-scale sand remobilization and injection, it is imperative that the full extent of remobilization and injection is realized at an early stage. Hence, we suggest that the North Sea experience should be borne in mind when exploring for oil and gas in similar settings elsewhere, including the prolific slope settings offshore northwest Europe, west Africa, and Brazil. A number of factors facilitate the formation of large-scale sandstone intrusions, including
deep-water sedimentary systems characterized by clean, unconsolidated sands encased in sealing mudstones
high sedimentation rates leading to disequilibrium compaction overpressures
hydrocarbon migration providing fluid buoyancy and overpressure
tectonic activity causing shaking and potential liquefaction of overpressured, unconsolidated sands
When it can be established that factors 1, 2 and/or 3, and 4 were present during the formation of a particular sedimentary section, it should be considered that sandstone intrusions may be a significant component of the local hydrocarbon system. Because these conditions would have been present in many basins around the world, we suggest that the Cretaceous–Cenozoic onshore California and the North Sea Paleogene may well be the first discovered examples in a long series of basins characterized by large-scale sandstone intrusions.
Large-scale sandstone intrusions are imperfectly imaged by even the best quality 3-D seismic data. Factors that may lead to improved imaging of complex geology, such as large-scale sandstone intrusions include high-density 3-D seismic acquisition and careful prestack depth migration. Seismic modeling of outcrop geometries may help define the limitations and possible solutions in seismic imaging.
Three main types of intrusions have been defined in the North Sea Paleogene:
winglike intrusions that are located at the edges of thick isolated sandstone bodies that may be of depositional or intrusive (sill) origin
conical intrusions that emanate upward from a central apex detached from their parent sandstone body
crestal intrusion complexes that constitute a stock work of relatively thin, but volumetrically significant, sandstone intrusions above more massive parent sandstones
The formation of the large-scale intrusions is linked with the occurrence of thick bodies of clean, unconsolidated sands encased in sealing mudstones subject to overpressure generation by disequilibrium compaction. Influx of hydrocarbons may have exacerbated the overpressures, whereas the actual triggering of sand intrusion is commonly attributed to seismic shaking.
The occurrence of large-scale sandstone intrusions may be significant to hydrocarbon exploration and production in several ways:
They constitute volumetrically significant reservoirs (intrusive and extrusive traps).
They impact reservoir quality, distribution, and connectivity.
They constitute efficient fluid-migration paths through hundreds of meters of mudstones.
They may be drilling hazards.
The North Sea experience shows that the recognition of large-scale remobilization and injection of sand in a deep-water clastic play can cause significant changes in reserves estimation and risking and has positive as well as negative consequences for hydrocarbon prospectivity. We argue that large-scale intrusions are unlikely to be unique to the North Sea Basin, as evidenced by their abundance in deep-water deposits onshore and, thus, predict their future discovery in deep-water basins around the world.
We acknowledge the support of the Injected Sands Consortium (2000–2002: ChevronTexaco, Enterprise Oil, Norsk Hydro, Shell, Statoil, and TotalFinaElf) and discussions with colleagues and collaborators at the universities of Aberdeen and Cardiff, in particular D. Duranti, S. Shoulders, and R. Jonk. Mads Huuse thanks F. Nadim and T. J. Kvalstad at the International Center for Geohazards, Norwegian Geotechnical Institute, for sharing their thoughts on deep liquefaction. Schlumberger and Landmark generously donated seismic interpretation and modeling software to the universities of Cardiff and Aberdeen, respectively. WesternGeco kindly allowed us to access and publish data from their 3D4C Chestnut survey. ENI Norge, Enterprise Oil, Norsk Hydro, Statoil, and Total are thanked for their cooperation and permission to publish data. D. Duranti and P. Nelson are thanked for helpful comments and suggestions to the manuscript.
Figures & Tables
Sand Injectites: Implications for Hydrocarbon Exploration and Production
Sand injectites are described in scientific literature as an increasingly common occurrence in hydrocarbon reservoirs, in particular in deep-water clastic systems, where they are known to influence reserves distribution and recovery. Seismically-detectable injected sand bodies constitute targets for exploration and development wells and, subseismic sand bodies provide excellent intra-reservoir flow units that create field-wide vertical communication through depositionally extensive, low-permeability units. As sand injectites form permeable conduits in otherwise low-permeability units they facilitate the expulsion of basinal fluids; hence they act both as a seal risk and mitigate timing and rate of hydrocarbon migration. Injected sand bodies form intrusive traps, which are distinct from structural or stratigraphic traps. Included in this publication are 10 chapters on subsurface examination of sand injectites, 1 chapter on theoretical considerations, and 13 outcrop analogs in reservoirs across the world. Captured in this volume is at least a taste of the global and stratigraphic distribution of sand injectites, and an attempt to introduce readers to sand injectites and their significance in the context of hydrocarbon exploration and production. The book is not intended as a complete review of the field-based literature, but emphasizes high quality case studies from the surface and subsurface. The geographic scope of the book is large, and illustrates the diversity of geological settings in which these fascinating and economically significant features are found.