Abstract
The Skarfjell oil and gas discovery, situated 50 km north of the Troll Field in the NE North Sea, was discovered by well 35/9-7 and was appraised by three additional wells operated by Wintershall, in the period 2012–14.
The Skarfjell discovery is an example of a combined structural/stratigraphic trap. The trap formed along the northern edge of a deep WNW–ESE-trending submarine canyon, which was created by Volgian erosion of intra-Heather, Oxfordian-aged sandstones and then infilled with Draupne Formation shales. This mud-filled canyon forms the top and side seal, with the bottom seal provided by Heather shales. The reservoir comprises mid-Oxfordian deep-water turbidites and sediment gravity flows, which formed in response to tectonic hinterland uplift and erosion of the basin margin, 10–20 km to the east.
The Skarfjell discovery contains light oil and gas, and may be subdivided into Skarfjell West, in which the main oil reservoir and gas cap have known contacts, and Skarfjell Southeast, which comprises thinner oil and gas reservoirs with slightly lower pressure and unknown hydrocarbon contacts.
The Upper Jurassic Draupne and Heather formations are excellent source rocks in the study area. They have generated large volumes of oil and gas reservoired in fields, and discoveries for which the dominant source rock and its maturity have been established by oil to source rock correlation and geochemical biomarker analysis. The Skarfjell fluids were expelled from mid-mature oil source rocks of mixed Heather and Draupne Formation origin.
The recoverable resources are estimated at between 9 and 16 million standard cubic metres (Sm3) of recoverable oil and condensate, and 4–6 billion Sm3 of recoverable gas. The Skarfjell discovery is currently in the pre-development phase and is expected to come on stream in 2021.
As exploration in the NE Norwegian North Sea is rather mature, in the sense that all the larger structural traps in the main plays have been drilled, recent exploration has focused on smaller structural traps, down-faulted traps and stratigraphic traps, often in areas with poor seismic imaging. The Skarfjell discovery is an example of a stratigraphic trap which was clearly identified on seismic data. The exploration well 35/9-7 was a commercial discovery, despite significant pre-drill concerns regarding the presence of reservoir and effectiveness of the trap.
The Skarfjell discovery well (35/9-7) and three additional appraisal wells (35/9-8, 35/10 S and 35/10 A) were drilled by Wintershall Norge AS in the period 2012–14, proving up the Skarfjell oil and gas discovery. All four wells encountered oil- and gas-bearing deep-marine, Upper Jurassic sandstones interbedded into the Heather Formation at two or three different levels.
The recoverable resources are estimated at between 9 and 16 million standard cubic metres (Sm3) of recoverable oil and condensate, and 4–6 billion Sm3 of recoverable gas. The Skarfjell discovery is currently in the pre-development phase and is expected to come on stream in 2021.
Geological setting
The Skarfjell discovery is located between the Gjøa and Vega fields, 50 km north of the Troll Field in the NE North Sea. The discovery is located mainly in Production Licence 418 (PL418) but extends into Production Licence 378 (Fig. 1).
Structural elements of the NE North Sea. The white line shows location of the seismic line in Figure 2. Field names are shown in white. Structural elements are from http://www.npd.no. The Skarfjell discovery is highlighted with the white frame.
The area is bordered to the east by the basin margin, comprising the Øygarden Fault Complex. Caledonian basement rocks of west Norway are exposed to the east of this fault complex. To the south is the Horda Platform, including the giant Troll Field, and to the west and north are the Magne Sub-basin (the deepest part of the North Viking Graben) and the Sogn Graben. These latter elements are separated by the Marflo Spur (see the structural elements map: Fig. 1).
Extensional faulting occurred in several tectonic phases, during the mid–late Jurassic (e.g. Beach et al. 1987; Giltner 1987; Badley et al. 1988; Blundell & Gibbs 1990; Ziegler 1990, 1992). This took place along large north–south- and NE–SW-trending normal faults, resulting in a horst and graben landscape dominated by eastward-tilted fault blocks, such as the Rygsteinen Ridge, the Vega Spur and the Afrodite fault block. These tilted blocks become deeper towards the west and north, as seen in the east–west seismic line shown in Figure 2.
Regional seismic east–west line through Skarfjell and Vega. Seismic horizons: 1, Post-glacial Angular Unconformity (PAU); 2, Top Hordaland Group; 3, Top Shetland Group; 4, Top Svarte Formation; 5, Top Cromer Knoll Group; 6, Top Åsgard Formation; 7, Base Cretaceous Unconformity (BCU); 8, Volgian Unconformity; 9, Intra-Heather; 10, Top Brent Group. TWT, two-way time (in seconds). Arbitrary line from PGS Mega-Merge. Published with permission from PGS.
The stratigraphy of the area is relatively complete in the western (basinal) part but the Triassic, Jurassic and Cretaceous intervals thin to the east. The Upper Triassic–Lower Jurassic Statfjord Formation directly overlies the Caledonian basement in wells in or close to the Øygarden Fault Complex (e.g. 36/7-1 and 36/7-2).
Rift-related differential subsidence of individual fault blocks during the late Jurassic (notably the early Oxfordian) was accompanied by deposition of anoxic shales, which form potential source rocks. In addition, rejuvenation of the hinterlands resulted in erosion/reworking of the shallow-marine Sognefjord Formation off of the Horda Platform and Øygarden Fault Zone, and caused deposition of coarse-grained deep-water sediment gravity flows in basin lows on the terrace areas (see Ravnås et al. 2000).
Syn-tectonic erosion in this area is recorded by subtle unconformities developed below the gravity-flow sections in the mid-Oxfordian, Kimmeridgian and Volgian. Locally, the depth of Volgian erosion was significant, especially close to the basin margin, where sub-marine erosion created a number of WNW–ESE-trending submarine canyons with deep truncation into the lower Callovian-aged part of the Heather Formation (Fig. 3). The lithostratigraphy of the area has been described by Vollset & Doré (1984) but no formal name was established for the Oxfordian gravity-flow succession, which forms the main reservoir interval in the Skarfjell area.
Late Jurassic stratigraphy of the NE North Sea. The Callovian–mid-Oxfordian stratigraphic interval, with the Skarfjell reservoirs Intra-Heather-1 and -2 (IH-1 and IH-2), is marked by the red frame. Colour code: yellow, shallow-marine sandstones; orange, deep-marine turbidites; light green, clay and siltstones of the Heather Formation; dark green, claystones of the Draupne Formation. SK marks the position of Skarfjell immediately NE of a sub-marine incision.
Exploration history
Hydrocarbon exploration in the Skarfjell area began with drilling of the 31/2-1 well in 1979, which discovered the giant Troll Field. Subsequent wells, drilled in 1980–81, discovered the Vega North Field (discovered by 35/8-1) and the Vega Sentral Field (discovered by 35/8-2). Later exploration drilling during the 1980s appraised these fields and discovered Fram and Gjøa; these four fields are producing today (Fig. 1). These fields are all traditional upthrown Jurassic fault blocks with structural closures and Middle Jurassic Brent or Upper Jurassic shallow-marine Sognefjord reservoirs. Recent exploration in the area has focused on smaller structural traps, downfaulted structural traps and stratigraphic traps.
The 35/9-7 well, which discovered the Skarfjell oil and gas accumulation, and three appraisal wells (35/9-8, 35/10 S and 35/10 A), were drilled in PL 418 by Wintershall Norge AS in the period 2012–14. All wells encountered oil- and gas-bearing deep-marine sandstones interbedded within the Heather Formation at two or three different levels.
Petroleum system overview
Hydrocarbon system
The area has a proven hydrocarbon system, consisting of Upper Jurassic source rocks, mid–late Jurassic primary reservoirs, Upper Jurassic and Cretaceous shales (forming the top seal), and mainly upthrown Jurassic fault blocks with structural closure. In more detail, the petroleum system can be characterized as follows.
Source rocks/hydrocarbon presence
Rich Upper Jurassic Heather and Draupne marine source rocks are mature for oil in the eastern part, and for gas in the deeper area towards the west and north. Hydrocarbon presence in the area (see Fig. 1) includes the following accumulations: oil with gas caps in the Gjøa, Titan, Skarfjell, Astero, Grosbeak and Fram discoveries; rich gas condensate in Aurora and the Vega fields; and dry gas in the Afrodite discovery (at the edge of the deep Magne Sub-basin).
Reservoirs
The middle Jurassic shallow-marine Brent Group has a typical gross thickness of 150–200 m and forms good quality reservoirs across the area. These typically thin and eventually shale out towards the NE in the study area.
The Upper Jurassic shallow-marine reservoirs in this area are dominated by sandstones of the Krossfjord (mainly Bathonioan), Fensfjord (mainly Callovian) and Sognefjord (Oxfordian) formations. These prograded generally westwards from the Norwegian mainland, across the Horda Platform.
Upper Jurassic deep-marine gravity-flow deposits were sourced during tectonic events that triggered gravitational instability of poorly lithified shallow-marine Sognefjord sands and sandstones. These sands were transported through canyons and deposited in relatively deep water with varying thicknesses of up to 200 m over large parts of the faulted terrace area (Fig. 4). These intra-Heather sandstones form the reservoir of the Astero, Aurora, Titan and Skarfjell discoveries.
In addition, some reservoir potential is present in the Lower Cretaceous Agat Formation along the Øygarden Fault Zone north of the Gjøa Field, and in Palaeogene sandstones.
Mid-Oxfordian depositional environments, showing the potential distribution of mid-Oxfordian deep-marine sandstones on the terrace area downdip of the transition zone and shallow-marine facies of the Sognefjord Formation along the basin margin. Brown arrows indicate the potential sediment entry points.
Trap types
The majority of the discoveries and fields in the Skarfjell area comprise tilted fault blocks with structural closure at the Base Draupne–Base Cretaceous unconformities, and are sealed by overlying Upper Jurassic and Cretaceous shales.
However, the Skarfjell discovery is a stratigraphic trap formed by updip erosional truncation of intra-Heather turbidites by WNW–ESE-trending, shale-filled submarine canyons.
Migration
Gas is believed to have migrated through the Brent aquifer on a semi-regional scale from deeper areas in the NW towards lower pressures to the SE, where hydrostatic pressure exists in the Gjøa, Fram and Troll fields. Locally sourced hydrocarbons (see the discussion below) were derived from oil-mature kitchen areas and reservoired in intra-Heather turbidites, resulting in slight overpressure. This overpressure may have prevented later gas flushing.
The Skarfjell trap
The Skarfjell discovery has a stratigraphic trap with a structural component. The discovery is located on the north–south-trending and generally eastward-dipping Rygsteinen Ridge (Figs 5 & 6). The main reservoir comprises mid-Oxfordian turbidites deposited during one of the main rifting phases in the northern North Sea.
Regional seismic section, showing the general eastwards tilt of the horst blocks, bounded by major westwards-dipping faults. The footwall scarp erosion of the Heather Formation in the western part of the Skarfjell high and the subsequent deposition of Draupne Formation shales is setting up the Skarfjell stratigraphic trap towards the west. Note the updip truncation of the mid-Oxfordian turbidites by the Volgian Unconformity (white arrow). Arbitrary line from PGS Mega-merge. Published with permission from PGS.
Regional seismic section along the north–south-trending Rygsteinen Ridge. The deep incision of the Volgian Unconformity into the Heather Formation south of the Skarfjell discovery is setting up a stratigraphic trap and separating Skarfjell from Grosbeak. Note the updip truncation of the mid-Oxfordian turbidites by the Volgian Unconformity (white arrow). Arbitrary line from PGS Mega-merge. Published with permission from PGS.
The Skarfjell stratigraphic trap formed after the main rifting event in the Middle Oxfordian, during latest Jurassic (intra-Volgian) hinterland uplift and incision of deep NW–SE-trending submarine canyons into the top of the Heather Formation (Fig. 7), mainly near the Øygarden Fault Complex and the eastern basin margin. The Skarfjell trap was formed along the northern edge of one of these submarine canyons, where intra-Heather sandstones were eroded below the Volgian Unconformity. During later Volgian subsidence, this valley was infilled with the Draupne Formation shale that forms the top and side seals. Later local and semi-regional tilt to the north and NE has formed a genuine stratigraphic trap, consisting of updip truncation of the intra-Heather sandstones towards the SE and a three-way dip slope towards the NE. Footwall erosion of the Heather Formation along the Kinna Fault and subsequent deposition of the Draupne Formation sets up the stratigraphic trap component towards the west. The bottom seal is provided by Heather Formation shales. The trap is able to hold a hydrocarbon column of more than 400 m.
Effect of the Volgian erosion (white areas) on the presence of the Oxfordian section of the Heather Formation (grey areas). Note that erosion of the Oxfordian section occurred on the uplifted footwall block of the Rygsteinen Ridge west of the hatched black line, as well as in elongate WNW–ESE-trending incisions, believed to have formed as submarine canyons close to the eastern basin margin. Note the incision south of the Skarfjell discovery (SF) that sets up the Skarfjell trap. The light grey polygons represent Upper Jurassic faults. Hydrocarbon fields are marked in dark grey.
Truncation of the mid-Oxfordian turbidites by the Volgian Unconformity at Skarfjell is clearly visible on seismic data (Fig. 8). The Volgian Unconformity/Base Draupne Formation is expressed as a distinct seismic reflector, as is the base of the mid-Oxfordian turbidites (Fig. 8). The Skarfjell trap is cut by numerous faults, which have varying offset along strike and also have strike-slip components. Despite the fact that most fault activity ceased during the latest Jurassic, some faults post-date the Volgian Unconformity and propagate into the Lower Cretaceous (Fig. 8). All of this fault activity is thought to predate the onset of hydrocarbon generation and migration, and apparently does not have any influence on trap integrity. Neogene and post-glacial (Pleistocene?) isostatic uplift of the Norwegian mainland (e.g. Coward et al. 2003) resulted in a further slight westwards tilt of the Skarfjell discovery, which was most probably charged by that time. However, there is no evidence of partial leakage (palaeocontact) of previously trapped hydrocarbons in the discovery. This implies that the trap has remained intact during the isostatic tilt.
Seismic section across the field, showing erosion of the Oxfordian turbidites towards the west. Note that most faults predate the Volgian Unconformity (red horizon), but some also seem to have younger movements. Note the thickening across the faults indicating syn-tectonic deposition, combined with erosion by the Volgian Unconformity. Horizon colours: green, BCU; red, Volgian Unconformity; yellow, Base Oxfordian turbidites (BOT); magenta, Top Brent Group. The white arrow highlights the truncation of the Oxfordian turbidites by the Volgian Unconformity. From seismic survey RD1201 PSDM.
In order to understand the trap configuration for Skarfjell, it is important to take a closer look at the nature of the Volgian Unconformity and the palaeorelief at the time of deposition of the Draupne shales, which provide the top and side seal for the Skarfjell discovery. During Volgian times, the Skarfjell area was an eastwards-dipping palaeohigh bounded by the Kinna Fault in the west, where relatively deep erosion of the Heather Formation took place on the highest part of the footwall (Figs 5 & 7). It is possible that this erosion occurred without any subaerial exposure. Along the southern limit of the Skarfjell trap, the Volgian Unconformity has formed a prominent WNW–ESE-trending (Figs 6 & 7) incision which seems to follow the trends of a group of NW–SE-trending faults, implying that changes in lithology/competence influenced the location and orientation of the submarine canyons.
A combination of detailed structural mapping and data from pressure analysis (see the section on ‘Hydrocarbons and field segmentation’ later in this paper) led to subdivision of the field into two compartments: Skarfjell West and Skarfjell Southeast (Fig. 9). As the pressure data suggest communication between the 35/9-8, 35/9-7 and 35/9-10 A wells, and as the throws of the smaller mapped faults are less than the reservoir thickness, there is no evidence to suggest further fault compartmentalization within Skarfjell West. The reservoir in Skarfjell West can be mapped with confidence based on seismic data. The reservoir thins into Skarfjell Southeast, where mapping confidence decreases.
Skarfjell Top IH-2 reservoir depth map in m TVDSS. Colour-coded map=above OWC (2783 m) in Skarfjell West. Red line, gas–oil contact (GOC) (2566 m); solid black line, possible pressure boundary between Skarfjell West and Skarfjell Southeast; dashed green line, possible GWC (2783 m) in Skarfjell Southeast; red line, GDT (gas down to) (2566 m) in Skarfjell Southeast. The dashed black line shows the well correlation in Figure 10; the dashed purple line shows the cross-section in Figure 17.
The Skarfjell reservoir
Intra-Heather-1 and the Fensfjord Formation
The oldest reservoir units in Skarfjell are relatively thin, gravity-flow sandstones informally referred to as the ‘Intra-Heather-1’ (IH-1) sands. These sandstones are of Bathonian age and were encountered in all four Skarfjell wells (Fig. 10). The gravity-flow sandstones that form the main reservoir in the Skarfjell Field are informally referred to as ‘Intra-Heather-2’ sands. In Skarfjell, these are of Early–Middle Oxfordian age. The underlying Early Oxfordian Heather mudrocks contain belemnites, and display generally intense bioturbation, dominated by Phycosiphon, Nereites, Schaubcylindrichnus and Teichichnus. These mudrocks are therefore thought to record deposition on a marine shelf, below storm wave-base. Wells 35/9-10 A and 35/9-10 S record progradation of wave-rippled shoreface sandstones that are extensively bioturbated by Asterosoma, deposited above fair-weather wave-base. The shoreface sands are thicker in well 35/9-10 S, and are therefore thought to have prograded largely from east to west and to be the most distal part of the Fensfjord Formation. The metre-scale shoreface sand bodies are seen in core and on wireline log data to have sharp, abrupt bases. This may be due to rapid progradation, perhaps forced by tectonic uplift of the shoreline (see Plint 1988). Break-up of the shelf in Early Oxfordian times was probably a result of more significant tectonic activity, generating sub-basins in which gravity-flows were deposited. The creation of these sub-basins appears to coincide with a change from bioturbated to laminated shales, which is interpreted to record reduced seafloor circulation within the deepening intra-shelf sub-basins.
Well correlation of the Skarfjell reservoirs, depth in m TVDSS, flattened on Top IH-1 reservoir. Horizons: green, Base Cretaceous Unconformity; purple, Volgian Unconformity; yellow, Top IH-2 reservoir; orange, Base IH-2 reservoir/Oxfordian Unconformity; brown, Top and Base of IH-1 reservoir. Reservoir intervals: (a) Intra-Heather-1 reservoir (IH-1) consisting of a thin Bathonian turbidite sequence; (b) Fensfjord Formation, consisting of distal shallow-marine sandstones and siltstones; (c) Intra–Heather-2 (IH-2), main Skarfjell reservoir of mid-Oxfordian age, consisting of Lower Sand Body (LSB) and Upper Channel-Fill (UCF). The location is shown by a stippled black line in Figure 9.
Lower Intra-Heather-2 (‘Lower Sand Body’)
This interval is 49.7 m thick in well 35/9-8, where it is similar in log/wireline facies to a slightly more amalgamated interval 26.8 m thick in well 35/9-7 (Fig. 10). Lithofacies correlation of the underlying mudrock intervals suggests approximately 1.5 m more erosion below the overlying sand body in 35/9-7 than in 35/9-8. This is consistent with 35/9-7 being located in a more proximal/axial position (in the south). Structureless, fine- to medium-grained sands interpreted to be high-concentration turbidites (e.g. Mulder & Alexander 2001; Talling et al. 2012) were passively dumped onto this basal erosion surface (Fig. 11). There are no associated basal lag deposits or tractional bedload facies (see Mayall et al. 2006; Sequeiros et al. 2010).
Core log and wireline log correlation of the Lower Sand Body. Note the general thickening in 35/9-8 downdip to the NE and the erosive base (local Oxfordian Unconformity) which has eroded slightly deeper in the 35/9-7 well to the south. For location see Figure 9. The lateral distance between the two wells is 1.8 km.
The sandstones are interpreted to be mainly structureless, high-concentration turbidites. They display coarse-tail grading, suggesting deposition from fully turbulent, depletive flows (Kneller 1995). Some beds grade up into minor ‘linked’ debrites (Haughton et al. 2003) and stratified, low-concentration, sandy turbidite caps (Fig. 12).
Core photographs from the Lower Sand Body of the 35/9-7 and 35/9-8 wells. (a) Erosional base of the Lower Sand Body; the white arrow shows the unbioturbated laminated facies that can be correlated between the two wells. (b, c) Dominant facies, high-concentration turbidites. (d) Middle part with the white arrow consists of carbonaceous pebbles, interpreted as a ‘linked’ debrite separating the high-concentration sandy turbidite below from the low-concentration turbidite above.
The sandstones are more strongly amalgamated (i.e. they display a greater proportion of sand-on-sand bed contacts) in 35/9-7 than in 35/9-8. This is consistent with a lower slope gradient in the north (resulting in less erosion at bed bases). Bed amalgamation increases upsection, suggesting that the system became more erosive and therefore ‘prograded’, probably in response to increasing slope gradients. The sands seem likely to form a broad slope apron that eventually shales out northwards. The presence of relatively thin and uncommon ‘linked debrites’ capping sandstone beds in both cored wells is thought to be characteristic of relatively proximal locations (Haughton et al. 2003, 2009; Hodgson 2009), implying that the shale-out was a significant distance to the north of the two well locations.
The observed vertical changes from structureless to stratified sandstones within this succession record reduced rates of sediment fall-out from suspension (e.g. Kneller 1995). The stratified sandstones commonly grade upwards from the tops of structureless intervals, rather than recording separate flow events. Some metre-scale intervals dominated by stratification are present, but these also include subordinate structureless and fluidized sandstone beds. The stratified sandstones are therefore not thought to represent a spatially separate, overbank facies. Instead, the Lower Sand Body is likely to comprise nested, sheet-like scour-fills that grade laterally from more to less structured sandstones, deposited as low-relief mounds of structureless sand draped by low-angle stratified sands. Some minor incision or mounding may have helped to concentrate the flows in low-lying areas, depositing structureless sand, lateral to stratified sand deposited on ‘highs’ or at low-relief channel margins. Very minor occurrences of muddy turbidites and linked debrites form small-scale and potentially discontinuous heterogeneity within this sand sheet/apron. Image-log data suggest low-angle depositional dips within the Lower Sand Body, inclined generally towards the east. Comparison with core data suggests that these dips reflect the orientation of low-angle inclined surfaces draping onto the margins of topography that is filled by the Lower Sand Body (see Kane et al. 2010).
This topography was probably generated by uplift of the Rygsteinen Ridge, which acted to confine the Lower Sand Body. A change in fault orientation (from NW and north trends to NE trends) is observed on seismic data further north. This appears to have guided the sand body into a generally NE-directed orientation in the vicinity of well 35/9-8. This change in orientation is consistent with the image-log dip data in that well.
Biostratigraphy suggests that the Lower Sand Body cuts down into the underlying Heather interval in 35/9-10 A (Fig. 11). The Lower Sand Body cored in 35/9-10 A comprises a 19.3 m-thick interval of amalgamated high-concentration sandy turbidites. These deposits dropped out of turbulent suspension from sandy gravity flows, depositing medium- to coarse-grained sands. Pebbles and cobbles of Heather mudstone present at the incised base of the interval may be derived from sub-marine erosion of the incised basal surface. Robust, abraded/broken, shallow-marine bivalve shells (?oysters) are common in some coarse-grained sandstones in the lower part of this interval. Similar bivalve shells are common in shallow-marine Sognefjord sandstones further east. This suggests reworking of unlithified shallow-marine sands and ?oyster shells possibly during input from a river in flood (see Mulder et al. 2003).
The equivalent interval in well 35/9-10 S comprises stratified fine- to medium-grained sandstones intercalated with thin layers of granules and small pebbles of lithic material, recording repeated bypassing events that deposited bedload gravel from tractional gravity flows. This suggests a relatively steep, bypassed slope. There is a break of slope seen on seismic to the north of this well, beyond which the Lower Sand Body present in wells 35/9-7 and 35/9-8 was deposited.
It seems likely that the coarser-grained sands in 35/9-10 A were deposited in a slope channel (e.g. Mayall et al. 2006), because they contain large clasts (cobbles and boulders) of Heather mudstone in core. This slope channel was likely to have been incised into the slope that was bypassed in 35/9-10 S, and passed northwards into thicker, finer, more widespread sands deposited to the north of the break of slope seen on seismic.
Upper Intra-Heather-2 (‘Upper Channel-Fill’)
The ‘Upper Channel-Fill’ is a heterolithic, pebbly succession, cored in wells 35/9-7 and 35/9-8, where it is up to 80 m thick (Fig. 13).
Core photographs from the Upper Channel-Fill of the 35/9-7 and 35/9-8 wells. (a, b) Slide deposit; the white arrow in (a) marks the base of the Upper Channel-Fill. (c, d) Pebbly lag deposits.
The basal part contains metre-scale, deformed blocks of shelfal mudstone, some of which are dated as earliest Middle Oxfordian. These blocks are interpreted to be slide blocks (olistoliths: e.g. Mulder & Alexander 2001; McHargue et al. 2011), which are interbedded with gravelly bedload deposits. The latter record tractional flow conditions (see Sequeiros et al. 2010).
The gravels locally contain (?shelf-derived) belemnite debris, and are thought to represent the matrix of slide deposits (olistostromes) derived from gravitational collapse of shelfal mud and sand. Dune-scale cross-stratification is locally developed in intercalated sandstones, indicating relatively confined flows that allowed dunes to form. Overall, this interval is interpreted to represent part of a channel-fill dominated by bank-collapse (slide) deposits and highly energetic tractional grain flows, as seen in many submarine slope channel-fills worldwide (e.g. McHargue et al. 2011).
Slumped sandstones are present, which in this model are thought to record the local collapse of channel-axis sands. The slide deposits, bedload gravels and locally derived slumped sandstones are interpreted to have formed in a confined slope channel setting. Apparent truncation of the Lower Sand Body underlying the channel-fill suggests that the channel was incised 15 m deeper in well 35/9-7 than in 35/9-8 (see Fig. 11). The incision of the channel probably records an increase in slope gradients due to tectonic activity.
The succession fines upwards overall, with preservation of in situ mudstones becoming more common. The succession is composed of up to eight 10 m-scale muddying- and fining-upwards sequences, each of which probably represents the fill of a separate channel. The cored facies (Figs 13 & 14) comprise:
shelf-derived slides, or mass-transport complexes, consisting of large, weakly deformed blocks of shelfal origin that slid into the slope channel;
bank-collapse conglomerates, comprising reworked overbank muds in a matrix that commonly comprises bedload sands and gravels;
bedload gravels, which contain extraformational clasts transported within tractional grain-flows;
high-concentration sandy turbidites, which were deposited from suspension and are likely to have been contained within channel axes;
overbank sands and muds, which record episodic low-concentration flows spilling beyond channel margins;
abandonment muds, which form a very minor component.
Core photographs from the Upper Channel-Fill of the 35/9-7 and 35/9-8 wells. (a) Thin heterolithic overbank deposits. (b) Bank-collapse deposits. (c) Banded silty sandstones; the white arrow is separating structureless high-concentration sandy turbidite below from low-concentration turbidites above.
The internal architecture of the channel-fill probably reflects the vertical stacking patterns seen in core. Figure 15 shows an attempt to correlate 10 m-scale packages in wells 35/9-7 and 35/9-8. Each package probably represents the fill of a channel, nested within the overall composite slope channel-fill (‘channel complex’ sensu McHargue et al. 2011), which is 80 m thick in well 35/9-8. This is similar to the major ‘third-order’ slope channels that form on passive continental margins (e.g. Mayall et al. 2006). Smaller-scale (‘fourth–fifth order’ sensu Mayall et al. 2006; ‘channel elements’ sensu McHargue et al. 2011) are probably stacked in an offset pattern within the master channel. Eight of these elements are present: each is 10–30 m thick. These are likely to be about 200–600 m wide, assuming an approximately 20:1 width/thickness ratio, as documented by McHargue et al. (2011) using seismic and well data, plus outcrops from the Magallanes Basin of Chile (Hubbard et al. 2008), and the Brushy Canyon in the USA (e.g. Pyles et al. 2010). Each of the 200 m-wide channel-fills is likely to pass from pebbly to sandy then heterolithic facies in a downstream direction, due to a downflow (northwards) reduction in slope gradient and thus an increase in the proportion of entrained sand that was deposited within each channel. Lateral changes of a similar nature can be anticipated towards the margins of the system, possibly accompanied by large-scale lateral accretion, offset stacking and incision of smaller-scale channels within the main channel belt (see Abreu et al. 2003; Mayall et al. 2006).
Core log and correlation of the Upper Channel-Fill. Note the erosive base, and the thickening of the upper and lower units in 35/9-8 ‘downdip’ to the NE. The 35/9-7 section is dominated by bedload gravels, whereas the downdip section is more varied with high-concentration turbidites, overbank muds and sands, as well as bank-collapse conglomerates. The lateral distance between the two wells is 1.8 km.
Biostratigraphic data indicate that the approximately 30 m-thick sandstone immediately overlying the cored interval in well 35/9-10 A is approximately time-equivalent to the ‘Upper Channel-Fill’ in the Skarfjell wells further north. Wireline log character suggests that it is generally sandier and less conglomeratic that the cored Upper Channel-Fill elsewhere, implying a less axial location in the system. These uncored sands are described as ‘fine to coarse grained’ on completion logs, which is probably too coarse for them to be overbank/levee deposits. They might represent in-channel sands that escaped from the channel during avulsion episodes.
The time-equivalent interval in well 35/9-10 S comprises pebbly sandstones containing lithic clasts as well as large mudclasts. The latter suggest incision into contemporaneous mudstones, followed by bank collapse. Rippled and laminated fine-grained sandstones at the top of the cored interval are low-concentration turbidites that may record overbank deposition. The interval is thought to record deposition within an incised channel, in a less axial position than the coarser-grained conglomerates cored in wells 35/9-7 and 35/9-8.
Figure 16 shows two sketch maps that illustrate the general depositional setting recorded by the Intra-Heather-2 units in the Skarfjell area as an overlay on the Base IH-2 structural depth map. Note the presence of mainly NW-trending faults in the south of the region and NE-trending faults in the north of the region. This fault pattern might have affected depositional topography, although the exact timing of fault activity has not been determined.
Depositional models for the IH-2 reservoir plotted on the Base IH-2 structural depth map. (a) Depositional model for the Lower Sand Body in 35/9-7, 35/9-8 and 35/9-10 A in yellow, deposited from the south in a possible confined fan that widens downdip further north. The 35/9-10 S sandstones are interpreted as turbidites with a westerly transport direction, deposited on a largely bypassed slope. (b) Depositional model for the Upper Channel-Fill in 35/9-7 and 35/9-8 (grey brown). The 35/9-10 S is likely to have been in a bypass area, whereas the 35/9-10 A stacked lobe facies may (yellow) belong to a different pulse of deposition.
Later truncation occurred at the intra-Volgian Unconformity, which locally cuts down into the top of the Skarfjell reservoir. The present extent of the Skarfjell reservoir sands onto the Rygsteinen Ridge has been modified by erosion at the Intra-Volgian Unconformity and the BCU. The shape of the sand bodies implied by Figure 16 is schematic; reservoir sands may extend beyond these limits until truncated by the overlying unconformities.
Hydrocarbons and field segmentation
The Skarfjell discovery contains light oil and gas (Figs 9 & 17). The discovery may be subdivided into two broad compartments along faults that are thought to create a pressure boundary between Skarfjell West and Skarfjell Southeast (as summarized earlier). The reservoir stratigraphy is shown in Figure 10.
North–south cross-section of the Skarfjell discovery showing both the IH-2 (main) reservoir and the IH-1 reservoir, with hydrocarbon fill shown by colour. The location is shown by the purple dashed line in Figure 9.
In Skarfjell West, the IH-2 reservoir interval is 55–130 m thick and contains oil in the deeper part (to the north), with a gas cap over the top of the structure to the south (wells 35/9-7, 35/9-8 and 35/9-10 A) (Fig. 9). The oil is saturated and in pressure communication with the gas cap. The pressures of the 35/9-7 and 35/9-8 line up on a common oil gradient in the main IH-2 reservoir in Skarfjell West, with the oil–water contact (OWC) at 2782.7 m TVDSS (true vertical depth subsea). The gas discovered in the main reservoir of the 35/10 A well is geochemically and compositionally consistent with the oil, and the gas–water contact (GWC) is deduced from pressure data at 2566.5 m TVDSS (Fig. 18). As the crest of the IH-2 is located at around 2350 m TVDSS, the Skarfjell West compartment has more than 400 m of hydrocarbon column in the main reservoir.
Skarfjell formation pressures.
In Skarfjell Southeast (35/9-10 S), gas was found in the IH-2 (main) reservoir interval, with a composition similar to the gas cap in Skarfjell West but with 3 bars lower pressure. In addition, the shallow-marine Fensfjord Formation, present in distal lower shoreface facies (Fig. 10), contains undersaturated oil in an oil-down-to (ODT) situation. The undersaturated Fensfjord oil has a slightly different composition and 3 bars lower pressure than that in the IH-2 in Skarfjell West. As no water pressure data are available for the Skarfjell East compartment (ODT in 35/9-10 S), the hydrocarbon contacts are unknown in this area. The IH-2 aquifer of Skarfjell West has approximately 12 bars of overpressure, which decreases downwards in the stratigraphy and along the spill route towards the SE. The pressure is 3 bars lower in the Fensfjord Formation in Skarfjell East (35/9-10 S), a further 4 bars lower in the 35/9-12 well, and has dropped further to hydrostatic pressure in the 35/12-5 well and further south and east. The Skarfjell East compartment could therefore have a deeper OWC than Skarfjell West.
Oil was also found in the thin IH-1 reservoir (Fig. 10) in all four Skarfjell wells with an ODT situation: that is, the hydrocarbon contact in this reservoir is not known. The hydrocarbon pressure in these thin sands appears to line up very close to the oil gradient in the main reservoir of Skarfjell West (Fig. 18). If the aquifer pressure of these thin IH-1 sands is lower than the overpressured IH-2, then the IH-1 sands may have a deeper OWC than in Skarfjell West.
The Skarfjell oil composition is slightly different in the west to east of the field and is slightly lighter in the deepest IH-1 reservoir than in the Upper IH-2 reservoir. The oil properties from Skarfjell West are summarized in Table 1.
Skarfjell West oil properties from the 35/9-7 pressure–volume–temperature (PVT) analysis
The reservoir fluid contains relatively small amounts of carbon dioxide (0.647–0.872 mol%) and minimal quantities (0.2–0.8 ppm) of hydrogen sulphide (35/9-8 drill stem test (DST)). The asphaltene content is 0.4–0.6 wt% and the wax content is 2.7–8.7 wt%.
Oil to source rock correlation
This section is based on detailed geochemical analysis of the Skarfjell discovery wells plus a large number of wells from other discoveries and fields in the Skarfjell area.
The Draupne and Heather formations are excellent source rocks in the Skarfjell area, where they have generated large volumes of oil and gas reservoired in fields and discoveries. Their typical initial source rock properties are summarized in Table 2.
Initial source rock properties of the Draupne and Heather formations in the Skarfjell area
The pristane/phytane ratio is a parameter characteristic for source rock environment redox potential (Volkman & Maxwell 1986), and the sterane ratio C27/C29 is a parameter characteristic for the depositional environment. These fluid ratios have been plotted and compared with Draupne and Heather Formation source rock samples (Fig. 19).
The source rock characteristic biomarker ratio pristane/phytane is plotted against sterane C27/C29 for fluids and source rock samples from the Draupne and Heather formations in the Skarfjell area.
A high yield of pentacyclic C29 sterane reflects terrestrial-dominated facies, while dominance of C27 sterane indicate a dominance of zooplankton and red algae (Peters & Moldowan 1993). The Draupne Formation samples mostly have pristane/phytane ratios below 2, while Heather Formation samples have pristane/phytane ratios above 2. The pristane/phytane ratios for the oils in Skarfjell are around 2, which is in the border area between the Draupne and Heather formations. The condensates in the Skarfjell discovery display slightly higher pristane/phytane ratios of about 2.5, which is most likely to be an effect of pristane enrichment relative to phytane, due to evaporative fractionation.
All the analysed wells with source rocks from the Draupne and Heather formations were drilled on structural highs; in basinal areas, however, more marine and better-preserved organic matter is expected, with sterane C27/C29 ratios closer to 2. Based on the data summarized in Figure 19, the Heather Formation is interpreted to be the main source of the fluids in Vega, Aurora, Grosbeak, Gjøa well 35/9-1, the Brent Group in wells 35//9-6 S, 35/11-5, and in two wells analysed from the Fram Field (35/11-7 and 35/11-10).
The sterane C27/C29 and pristane/phytane ratios of the analysed oils in Astero and of the fluids in Upper Jurassic reservoirs in 35/9-2 (Gjøa), 35/11-5, and 35/11-8 S, 35/11-9, 35/11-10 (Fram Field) indicate that the source rock is mainly from the Draupne Formation. Hence, there are several accumulations (the Skarfjell Field, well 35/11-4, 35/11-7 (Fram) and the intra-Heather sandstone in well 35/9-6 S), which are interpreted to have a mixed petroleum source, with contributions from the Draupne, Heather, and, perhaps, also from Middle and Lower Jurassic source rocks.
The δ13C isotope values of saturated and aromatic fractions have been cross-plotted for fluids and source rocks in the study area (see Fig. 20). The saturated fraction of the Draupne Formation kerogen is characterized by δ13C isotope values between −28.5 and −32, whilst the aromatic fraction kerogen displays δ13C isotope values between −26.5 and −32. The corresponding δ13C isotope values for the saturated Heather Formation kerogen fraction are between −29.0 and −26, and the aromatic fraction displays δ13C isotope values of between −28 and −24. There is some isotope value overlap for the Draupne and Heather source rocks but the isotope values are generally characteristic of the origin of the fluids. Most of the fluid samples display isotope values between pure Draupne and pure Heather compositions, and within this group there are some clear trends.
Carbon isotope values for saturated and aromatic fractions for fluids and source rocks samples from the Draupne and Heather formations in the Skarfjell area.
The δ13C isotope values for fluids analysed from the Vega, Aurora, Brent Group in well 35/11-5, Brent Group in Fram, Dunlin Group in well 35/9-1 and Grosbeak correspond largely to the δ13C isotope values characteristic of the Heather Formation. In contrast, the fluids in Oxfordian sandstone reservoirs in wells 35/9-2, 35/11-5 and 35/11-8 S correlate mainly to the Draupne Formation.
A similar trend is seen in the source rock characteristic parameters, pristane/n-C17 ratios and phytane/n-C18 ratios (see Fig. 21). Most of the fluid samples plot in the area for mixed organic sources of marine and terrestrial organic matter. The fluids from the Brent Group in well 35/9-6 S (Titan), well 35/11-5, Vega and Grosbeak are thought to have been derived from terrestrial-dominated source rocks, with values that correlate mainly to the Heather Formation. In contrast, the fluids from the intra-Heather sandstones in wells 35/9-2 and 35/11-5, in Astero, and in Fram wells 35/11-8 S, 35/11-9 and 35/11-10 correlate mainly with the Draupne Formation.
Cross-plot of pristane/n-C17 against phytane/n-C18 ratios for fluids from wells and source rock samples from the Draupne and Heather formations in the Skarfjell area.
A wide range in maturity of the fluids is indicated by the absolute values of the pristane/n-C17 and phytane/n-C18 ratios, as indicated in Figure 21. The sterane ratio %20S [100×29ααS/(29ααS+29ααR)] appears not to have reached the endpoint of isomerization equilibrium for all of the fluids as show in Figure 22, where the maturity sensitive triaromatic sterane ratio 20TA/(20TA+28TA) is plotted against the sterane ratio %20S for fluids samples across the study area. The C26–C28 triaromatic steranes are less stable than the C20 and C21 triaromatic values. Immature sediments therefore display triaromatic sterane ratio 20TA/(20TA+28TA) values of between 0.1 and 0.2; this increases up to 1.0 for very mature sediments.
Cross-plot of the maturity sensitive triaromatic sterane ratio 20TA/(20TA+28TA) and sterane ratio %20S (29ααS/(29ααS/29ααR) for oils and source rocks samples from the Draupne and Heather formations in the Skarfjell area.
The fluids in Vega, Titan, Aurora, Astero and well 35/9-1 in Gjøa were all charged from very mature source rocks, whereas the fluids in Grosbeak and the oil in Fensfjord in 35/9-2 in the Gjøa Field were derived from early–mid oil-mature source rocks. The Skarfjell fluids were expelled from mid oil-mature source rocks. The main discoveries and the dominant source rocks to the fluids present are listed in Table 3 and shown in Figure 23.
Overview of the fields and discoveries in the Skarfjell area with interpreted source rocks to the fluids.
Fields and discoveries in the Skarfjell area with interpreted source rocks to the fluids
Resources and development
The recoverable resources are estimated between 9 and 16 million Sm3 of recoverable oil and condensate, and 4–6 billion Sm3 of recoverable gas. The Skarfjell discovery is currently in the pre-development phase and is expected to come on stream in 2021.
Conclusions
Stratigraphic traps are generally thought to be very risky. However, such broad generalizations should be avoided as there are many different types of stratigraphic traps with very varied risks. The degree of risk must be assessed carefully for each individual prospect under evaluation.
In Skarfjell, the base-of-seal, trapping surface could be clearly mapped. The reservoir sands were very difficult to predict, both in terms of presence and quality, but were expected to come from the east into the tilted trap and therefore not to pose an updip sealing risk. The main trap risk was considered to be the possible presence of thief zones in the incision fill.
By describing the elements of the Skarfjell discovery and stratigraphic trap, we hope to have contributed to a more detailed understanding of stratigraphic traps elsewhere.
Acknowledgments
The authors would like to acknowledge all the Wintershall staff that worked on the Skarfjell discovery including: Michael Zechner, Bernhard Frey, Marius Brundiers, Geert Strik, Tore Herrevold and several others. We furthermore want to thank Wintershall and the Skarfjell partners: Bayerngas Norge AS, Capricorn Norge AS, Edison Norge AS, DEA Norge AS and Repsol Exploration Norge AS for permission to publish the paper.
- © 2018 Petroleum Geology Conferences Ltd. Published by the Geological Society, London. All rights reserved
References
The 35/9-7 Skarfjell discovery: a genuine stratigraphic trap, NE North Sea, Norway
