Abstract
The Balder and Jotun fields are located on the western flank of the Utsira High, close to the eastern pinch-out of a Tertiary submarine fan system. Although similar in many aspects, the fields display important depositional, structural and stratigraphic differences which influence the choice of development strategy. The subsurface development strategy for the fields has been designed to optimize oil capture and minimize risk based on the interpreted reservoir geology. Therefore, differences between the exploration histories and reservoir geology are reflected in the development strategy of the two fields.
Both fields comprise Tertiary reservoir sands shed from the East Shetland Platform and transported across the Viking Graben area onto the Utsira High by sandy debris flows and turbidites. These distal gravity flow deposits display both thin-bedded sands and thicker more massive sandstones (>100 m). In the Balder Field, an intricate interaction between deposition and soft sediment deformation processes generated a complex network of reservoir compartments with common fluid contacts. In the Jotun Field, the oil–water contact is also common between all three structures, but a gas cap is restricted to one of the structures.
Introduction
The objective of this paper is to compare and contrast the Tertiary Balder and Jotun oil fields with respect to reservoir geology, and to illustrate how knowledge from the reservoir geology and field histories have influenced development strategy.
The Balder and Jotun fields are located about 190 km west of Stavanger, in the Norwegian sector of the North Sea. The Balder Field is located on production licences PL 001 and PL 028 on Block 25/11 and 25/10. Jotun is located 30 km to the north of the Balder Field on Blocks 25/8 and 25/7 ( Fig. 1). The fields are similar in size, with both expected to recover about 200 million barrels of oil.
North Sea structural elements and Paleocene deposition. Jotun and Balder fields are located on the western margin of the Utsira High, near the eastern limit of sand deposition. The Paleocene and Eocene sands were derived from the East Shetland Platform 50–100 km to the WNW.
Field History and Development
Balder
The Balder Field is located on licence PL 001 awarded to Esso in 1967 with a 100% working interest. The first well in the area, 25/11-1 (1969), had shows in thin early Eocene sands and the first litres of oil recovered from the Norwegian continental shelf were brought to the surface by a formation interval test (FIT). Five years later, the seventh exploration well discovered the field by demonstrating the potential of a commercial accumulation. The Balder Field was discovered by well 25/11-5 encountering a 25 m oil column in the Paleocene. The reservoir geology proved to be very complex and large uncertainties in predictability were experienced in the appraisal phase. Although the discovery by the 25/11-5 well was in 1974, it took a total of 18 exploration and appraisal wells, a five-month production test (1991) and two seismic 3D surveys (1979 (1988) before it was decided to develop the field. A gas cap was discovered in the last appraisal well 25/11-19S (1995). The first development wells proved the gas–oil contact to be a common field-wide contact. In 1996 to 1998 the Phase 1 development wells were predrilled from four sub-sea templates. Production start-up commenced on the 30 September, 1999, from the Balder FPU (floating production unit). The Balder Phase 2 development drilling started during the winter of 2000.
The initial development included 14 wells; seven horizontal producers, three vertical producers, two water injectors, a gas injector for associated gas and a water-source well ( Fig. 2). Of the three vertical producers, two have been replaced by horizontal wells while one has been abandoned due to high water-cut.
Balder Field development.
The Balder Field is a cluster of small Paleocene/early Eocene oil accumulations sharing a common oil–water contact and a field-wide gas cap. The field’s complex geology, combined with limitations in data quality, led to the proposal of different geological models and was one of the main reasons for the 30-year long exploration and appraisal phase ( Skjold 1980; Sarg & Skjold 1982; Hanslien 1987; Jenssen et al. (1993)). Uncertainties related to the complex geology also led to the implementation of a phased development strategy. Phase 1 included the six main mounds and a thin early Eocene sand accumulation to the northeast (E2). Balder Phase 2 drilling includes infill drilling of some Phase 1 mounds combined with nearby Paleocene mounds and early Eocene accumulations as part of the Ringhorne development project ( Fig. 2).
Jotun
The Jotun Field consists of three Paleocene structures produced through a wellhead platform and a floating production ship (Jotun A FPSO). The initial development drilling was completed in 2000 and infill drilling commenced during the summer of 2002.
The Jotun Field history is quite different from Balder as it is an example of a ‘fast track’ development that took only five years from discovery in 1994 to first oil in the fourth quarter of 1999. The field is mainly located on production licence PL 027B in Block 25/8 where Esso and Enterprise hold a 50% working interest each ( Fig. 3). A minor part of the field stretches westward onto licence PL 103 on Block 25/7, currently held by Conoco, DNO and Petoro. The Jotun Field was discovered by well 25/11-5S penetrating an oil column in the Elli structure in 1994. A 3D seismic survey was shot while the discovery well was being drilled. In 1995 well 25/7-3 proved oil in Elli South. The same year, 25/8-8S – with its side-tracks 8A and 8B – demonstrated that the Tau structure contains both oil and a gas cap and that the sand pinches out to the east ( Fig. 3). A field-wide oil–water contact at 2091 m TVDSS was defined from the exploration and appraisal wells. The appraisal programme was carefully designed from the 3D seismic interpretation and available wells in the area. The 25/8-8 well with its three deviated penetrations of the reservoir interval was planned to complete the appraisal drilling costeffectively while in parallel planning the field development. A total of four wells and three side-tracks were drilled pre-development. The field-wide contact was defined through averaging contacts observed in individual wells. When looked at in detail it seems clear that there are a few metres difference in the oil–water contacts between the three structures. In addition there is clear evidence of segmentation in Tau, with a deeper contact in the south (2094.4 m TVDSS defined by well 25/8-8B) and a shallower contact in the northern 25/8-B12 area (OWC 2088 m TVDSS in B12A and WUT 2076 m TVDSS in B12 toe pilot).
Jotun Field development.
The initial development plan comprised of 14 wells: eleven horizontal oil producers and three water injectors. Based on reservoir simulations and early production data the programme was altered to include three more producers; only one injector was drilled in this initial phase. Of the initial 14 wells, 12 are completed as horizontal producers, one as a vertical water injector and one as a water disposal well.
Regional Setting and Stratigraphic Framework
The Balder and Jotun fields are located on the northwestern flank of the Utsira High, which is a basement high along the eastern flank of the Viking Graben ( Fig. 1). The Utsira High was a positive feature as early as the Permian. When the late Jurassic block faulting and extensional structuring ceased, the North Sea basin became the site of passive subsidence during the Cretaceous and Tertiary. This created a north–south trending Paleocene basin, bounded by the East Shetland Platform to the west and the Norwegian shelf to the east. The Balder and Jotun fields lie near the centre of this basin but at the depositional edge of the sand system. Tectonic uplift of the Atlantic rift margin, combined with falls in eustatic sea-level during the Early Tertiary, caused erosion of the East Shetland Platform, and large volumes of sediment were deposited into the basin. Submarine fan complexes were derived from these shelf sands, shed from the East Shetland Platform and transported by gravity flows into the basin. The Jotun and Balder fields are both located close to the eastern termination of the same overall fan complex, with deposition across the Viking Graben and onto the Utsira High ( Fig. 1). Although these deep-marine sands are deposited in the distal portions of the fan system, thick massive sandstones (>100 m) were deposited mainly by debris flows but in part also by high density turbidites. Thinner turbiditic deposits down to millimetre-scale beds/laminae are also found in both fields.
The reservoir intervals for both fields are of Tertiary age and belong to the Rogaland Group ( Fig. 4). The Balder Field is comprised of three main reservoirs, the Paleocene Heimdal and Hermod formations and the early Eocene Balder Formation ( Figs 4 and 5). Neither biostratigraphy, seismic data, well logs or other independent correlation techniques allows for additional high confidence stratigraphic breakdown of the Heimdal Formation reservoir. Well-log correlations have been used to subdivide the Hermod and Balder Formation into subzones. The subzonation is not essential to the understanding of this paper and hence it is not discussed further. The reservoir of the Jotun Field is entirely within the Heimdal Formation and four subzones have been defined based on seismic data and well logs ( Figs 4 and 6).
Generalized stratigraphic column, composite logs and seismic horizons.
Balder Field structural/stratigraphic well log cross-section.
Jotun Field structural/stratigraphic well log cross-section.
Reservoir Mapping and Geological Model
Seismic and trap definition
In general, the seismic frequency is about the same for the two fields, with both seismic surveys in the 20–50 Hz data range. However, physical properties of the reservoir rocks, such as density contrasts and geometry, lead to significant differences in actual resolution between the two fields. The Top Chalk can be correlated with confidence across both field areas as elsewhere in the North Sea, and can be used to map palaeo-structure and fault trends which influence earliest Paleocene sedimentation in the area.
In the Balder Field area the Base Balder Formation (Base Tuff) marker can also be picked with a high degree of confidence throughout the survey area. The Top Balder Formation marker can be mapped with good confidence ( Figs 4 and 7). Top reservoir is only directly mappable on seismic data in parts of the field. Therefore, for other areas tops of reservoirs are mapped by isochoring up and down from the seismically defined surfaces. The Base Hermod Formation can be mapped in areas where it is thick (>20 m) and water bearing. A direct hydrocarbon indicator (DHI) from the oil–water contact can be mapped within mounds of massive thick sands. Increased amplitude response from hydrocarbon-bearing sands in the Balder Formation is also utilized in mapping the oil reservoirs ( Fig. 7). A detailed discussion of the Balder Field seismic mapping techniques and data quality of the two 3D surveys is given by Rye-Larsen 1994. The 1988 seismic survey was reprocessed in 1990. Several data cubes were generated (e.g. far offset, near offset, impedance etc.) and utilized to define the complex internal reservoir architecture segmenting the field into a cluster of accumulations. The impedance cube (CSSI, constrained sparse spike inversion) has proven to be very useful in reservoir definition, especially in the Early Eocene Balder Formation, detecting hydrocarbon sands down to 5–10 m in thickness. Thin-bedded reservoir identification and characterization through combining core, well log and seismic analysis is discussed in Haynes et al. (2000).
Balder Field random seismic section.
For the Jotun Field, top reservoir is mapped directly by the Top Heimdal marker. This seismic event is tied and correlates with well data. The Base Upper Heimdal is mapped with good confidence throughout the field and an intra-Upper Heimdal marker can be identified in some areas. These internal reflectors help in defining the reservoir geometry ( Figs 4 and 8). A DHI, corresponding to the oil–water contact, is present and can be mapped across most of the field, especially over the Elli structure. Phase reversal at the Top Heimdal marker was noted along the crest of the Tau structure. This has been confirmed by drilling to be a result of a gas cap in the Tau structure ( Figs 3 and 8). Multiple seismic cubes have been generated from the Jotun 3D seismic volume. These were used separately and in combination, for attribute studies utilized in reservoir mapping.
Jotun Field random seismic section.
The structure maps of the two fields generated from the depth-converted seismic interpretation of Base Balder and Top Heimdal markers illustrate the differences in the overall reservoir geometry between the fields ( Figs 9 and 10). The Balder Field is complex, with both steep flanked and lower relief multi-orientated Paleocene mounds. The intervening lows are filled with a heterolithic sediment mixture of tuff, sand and shale from the Early Eocene Balder Formation. Slump faults can be mapped along the mound flanks, but even if these are restored and the sediments are decompacted, the flank dips are still high, in the 8–10°range. The Balder Field can be described as a cluster of small hydrocarbon accumulations. This cluster is referred to as a field because of the common fluid contacts exhibited across the Balder Field and adjacent areas. Lack of a structural spill point indicates a stratigraphic component to the trapping mechanism. The Jotun Field is structurally simpler than the Balder complex and comprises three structures connected through a common saddle area ( Fig. 10). Elli and Elli South both display four-way structural closure, while the eastern Tau structure is a combination structural and stratigraphic trap. The Tau accumulation is limited to the east and southeast by the sand pinch-out of the fan system. A complicating factor for Jotun is that Tau has a gas cap not seen at Elli even though the crest at Elli is above the gas–oil contact observed at Tau; all three structures have the oil–water contact at about the same level (2091 m TVDSS). This may indicate lack of effective communication between the two structures, but it can also be explained in terms of hydrocarbon migration. The migration path to the structures may be through an area where the Chalk is eroded south of Tau. The Tau structure will in this case be filled first and spill to the western structures. In this scenario Tau is underfilled with respect to gas and therefore the gas never migrated to the Elli structure.
Balder Field base Balder Formation depth structure map, illustrating the irregular orientation and size of the Paleocene mounds with steep faulted flanks. Low areas shown in green have an irregular fill of Balder Formation sands, shale and tuff.
Jotun Field top Heimdal Formation depth structure map.
Reservoir stratigraphy
The Heimdal Formation reservoir at Jotun has been divided into four zones based on core descriptions and log correlation ( Figs 4 and 6). These subzones have been used to define reservoir geometry for 3D geological modelling and reservoir simulation to optimally position the development wells. The general stratigraphy of the Heimdal, Lista, Sele and Balder formations in the Jotun wells has been confirmed through palaeontology studies.
The reservoir stratigraphy of the Balder Field has been divided into the Heimdal, Hermod and Balder formations. The formations are distinguished by independent techniques such as well logs, shale colours, geochemistry and high resolution palynology. The Balder Formation (Palynozone III) is split into three subzones, while the Sele/Hermod interval (Palynozone II) is subdivided in two subzones based on well-log correlation. Palynology supports the division of Zone II, resulting in the Hermod sands being subdivided into upper and lower units. The Lower Hermod sand seems to be thicker and more widespread than the Upper. Several attempts have been made to divide the Palynozone IB Heimdal/Lista section into sub-zones through the years. However, any IB subzonation is regarded to be of low confidence as new wells commonly have generated data contradicting existing subdivisions of the zone. Hence, subdivision of the Zone IB is currently not used for any practical purposes in developing and producing the zone.
Depositional facies
A set of six generic facies is used to describe the reservoir sections in both fields ( Fig. 11): Facies 1, Sandy Debrite; Facies 2, Sandy Turbidite; Facies 3, Muddy Debrite; Facies 4, Turbiditic Shale; Facies 5, Hemipelagic Shale; Facies 6, Injectite.
Pictures of core samples: (a) Sandy Debrite with dish structures; (b) Sandy Debrite, fine to medium grained, with shale clast; (c) Sandy Turbidite, very fine to medium grained; (d) Muddy Debrite, very fine to medium grained, with shale clasts; (e, f) Turbiditic Shale; (g) Hemipelagic Shale; (h) Injectite with shale breccia; (i) Injectite with conjugate microfractures.
Sandy Debrite.
This is a term used to describe sands deposited by debris flows. These sand deposits are typically thick (several metres) with both upper and lower boundaries, sharp and conformable to the over- and underlying beds. Internal grain organization is referred to as random in that these beds lack grading, bed-sets or lamination ( Figs 11a and 11b). Typically, the beds lack primary sedimentary structures and display a homogeneous structureless distribution of sand grains. Coarser grains are randomly distributed from top to base, ‘floating’ in a fine- to medium-grained sand host interval which is moderately well sorted. Dish, pillar and other water escape structures are common in these deposits ( Fig. 11a). Amalgamation surfaces, occasionally observed in thicker units, indicate individual beds to generally be a few metres thick. The content of finer grain fractions, such as silt and clay, is normally less than 2%. Very faint horizontal thin beds/laminae, interpreted to have been generated from shear planes within the flow units, are observed in a few beds.
Sandy Turbidite.
These beds differ from the sandy debrite beds in that they are commonly thinner and display strong internal organization, such as graded beds and laminae ( Fig. 11c). The sandy turbidite beds are typically less than a metre to a few metres thick and commonly grade up into one of the shale facies. Evidence of erosion by truncation of underlying shale laminae is often observed at the basal bed boundaries. The Bouma sequence terminology (Ta–Td) is used to further define these beds in core descriptions. Ta and Tb units are commonly observed, whereas the rippled Tc facies is only occasionally present. Te and, occasionally, Td units are also observed in association with the sandy turbidites, and are often inter-beds rather than part of continuous fining-upward sequences. Laminae or horizons of imbricated shale clasts are commonly observed in Td intervals of beds with sharp upper and lower contacts.
Muddy Debrite.
This term is used to describe sand beds with high clay (mud) content. These beds contain poorly sorted very fine to fine sand with common coarser grains, including rounded to angular clay clasts. The beds often show clear evidence of soft sediment flow structures ( Fig. 11d). Larger clay clasts often occur dispersed throughout the beds and are typically deformed together with the beds’ flow structure. As for the sandy debrite the upper and lower boundaries are sharp. Occasionally, terrestrially derived organic material, in the form of coal and other plant fragments, is evenly distributed throughout the muddy debrite beds.
Turbiditic Shale.
This is the term used to describe shale beds containing thinly laminated turbidites of very fine sand and silt. The thin turbiditic beds are of a centimetre to millimetre scale and commonly occur as lamina rather than graded beds. The turbidites are typically either found at rhythmical intervals of a few to tens of centimetres or the entire bed is finely laminated by very thin turbiditic deposits ( Figs 11e and 11f). Bioturbation is minor, if present at all, and the turbiditic shale facies commonly contains terrestrial organic material. The Sele Formation is a good example of finely laminated organic-rich turbiditic shale ( Fig. 11f).
Hemipelagic Shale.
This is a mudstone with low silt content and a high degree of bioturbation ( Fig. 11g). The hemipelagic shale commonly has a waxy appearance and is non-fissile, blocky, with mottled colouring of greenish-grey, dark and light grey and reddish-brown. Organic content is very low. A diverse trace fossil assemblage is normally present. Among the most common burrows are Zoophycos and Condrites. Occurrences of these and other horizontal detritus-feeding forms are indicative of a quiet, deep-water environment.
Injectite.
This is the term commonly used to describe sand injection dyke and sill complexes. Sand injections are recognized in cores based on the sharp cross-cutting stratal relationships at upper and lower contacts with the host shale ( Figs 11h and 11i). Commonly the sand injections contain shale clasts from the host rock and can be described as a shale breccia ( Fig. 11h). Injections parallel to bedding are also observed. They are normally defined from core where they contain shale breccia, wedge out or are part of a larger injection complex. Sand injecting into sand is rare but has also been observed in cores. The grain size and reservoir quality of injections is very similar to the massive unconsolidated sands from which the injections originated.
Stacking pattern, reservoir architecture and geometry
Channel features with compensational stacking patterns have been observed in the Jotun Field by seismic stratigraphic interpretation. Core and well log observations, combined with seismic channel facies and isochore trends, show that sandy debrites and turbidites dominated the deposition along the channel axis. The term ‘channel’ is here used to refer to an elongated area of preferred deposition, and is likely to represent a confinement rather than an erosional channel feature. The channel axis consists of massive sand deposits which are compensationally amalgamated into elongated sandbodies with little or no internal shale beds ( Fig. 12a). Away from the channel axis and especially towards the eastern pinch-out, turbidites become more common and dominated the deposition. The change in facies is also reflected by thinning of the isochore away from the channel axis ( Fig. 12b). Elli and Elli South are homogeneous and have a high percentage of amalgamated sandy debrite facies, but sandy turbidite beds also occur. Turbiditic shales and occasionally hemipelagic shales separate the channel axis sandbodies.
Jotun Field channelized depositional pattern.
The Tau structure contains a more heterogeneous set of depositional facies, both in a vertical and lateral sense. The turbiditic facies in the Tau region produced a layered sheet-like reservoir sequence where sand continuity and reservoir connectivity is better defined laterally than vertically ( Fig. 6). Based on core descriptions, it is interpreted that the internal sand geometry in Elli and Elli South is mounded and that the connectivity is uniform and excellent, both vertically and laterally, with little or no baffes of shale within the reservoir section ( Figs 8 and 12). The exception to this is Zone 4 at Elli South in the sand injection and slump complex containing most of the oil in the structure. The sand injection complex, combined with fault juxtaposition, creates connectivity through the hemipelagic shale separating Zone 4 from the underlying massive Sandy Debrite section in Zone 3. Minor sand injections are also observed in Elli and Tau but these are volumetrically insignificant and likely to play only a minor role in the overall reservoir characterization. The turbiditic shale facies, associated with larger channel axis sandbodies can act as local baffes, whereas the hemipelagic shales, which in part can be correlated across the entire field, may act as barriers ( Figs 11c, 11e, 11g and 12). However, small faults (5–10 m) are mapped with a fairly high frequency across the field, which could create additional cross stratal communication through fault juxtaposition.
The three Balder Field reservoir sequences, Heimdal Formation (Zone IB), Hermod Formation (Zone II) and Balder Formation sand (Zone III), also display compensational stacking patterns where sands occupy the palaeo-lows during deposition. The Ty Formation (Zone IA2) contains the first sand deposited in the area which partly infills the pre-existing Chalk topography ( Fig. 13a). The amalgamated thick sandy debrites of the Heimdal Formation (Zone IB) partly fill the remaining low relief of the Chalk palaeo-structure and also produced a widespread constructural mounded sand deposit ( Fig. 13b). The Lista Formation hemipelagic shale and the Sele Formation turbiditic shale were deposited and draped over the Heimdal Formation mounds before the Zone II Hermod Formation sand was deposited in lows and along flanks of the pre-existing Zone IB topography ( Fig. 13b). The lower relief Hermod Formation mounds contain amalgamated sandy debrites and sandy turbidites. Finely laminated turbiditic shales of the Sele Formation are locally present and probably form flow baffes or potential barriers within Zone II mounds. Isostatic subsidence along the Viking Graben and associated westward tilting in the Balder Field area during the early Eocene caused sliding, slumping and sand remobilization in these Paleocene deposits. This enhanced the mounded geometry and sand injection complexes formed above the mounds. The Balder Formation thickens into the Paleocene lows and sand is irregularly distributed within the lows ( Fig. 13c). This indicates that the slump and slide deformational process enhancing the Base Balder topography occurs contemporaneously with Balder Formation deposition. It is interpreted that the deformation started during the late Paleocene (Zone II). Seismic onlap patterns demonstrate that it continued post-Balder Formation in the northwest, 25/10-3 area ( Fig. 13c). Observed circular and oval depressions in the Base Balder Formation structural surface were likely to have been caused by sand evacuation ( Figs 9 and 13c). The Balder Formation Mound 4 (M4) in the central low contains massive sand of amalgamated Sandy Debrite origin ( Figs 1 and 9). Outside of this mounded area the dominant facies in the Balder Formation is turbiditic and both sandy turbidites and turbiditic shale are found interbedded with the tuff beds. The basal part of the Balder Formation is often referred to as ‘the massive tuff’ and consists almost entirely of tuffaceous beds with sparse thin shale interbeds of hemipelagic origin. The basal part of the graded tuff beds is coarser and almost without exception calcite cemented. The ‘massive tuff’ is a barrier to vertical flow. This hydrocarbon seal is breached both by faulting and sand injections. Detailed attributes of the Balder Field depositional facies and reservoir geometry are further described in Jenssen et al. (1993).
Balder Field geological model.
Complex reservoir architecture of submarine fan depositional facies resulting from deformation by sand remobilization, comprising of slumps, slides and sand injections is not unique to the Balder and Jotun fields. Similar reservoir geometry and architecture has been reported from other North Sea fields such as Grane, Gryphon, Forth/Harding, Alba and in outcrop sections ( Newsom 1903; Newman et al. 1993; Dixon et al. 1995; Parize et al. 1997; Lonergan & Cartwright 1999; Lonergan et al. 2000).
Reservoir characteristics
Both fields have excellent reservoir characteristics in terms of porosity, permeability, reservoir continuity and connectivity ( Table 1). The lower average porosity and permeability in Jotun when compared to Balder is mainly a result of the additional compaction (300 m thicker overburden). Massive sand units of debritic origin display only minor variation in facies and have uniform distribution of reservoir properties due to the homogeneous nature of the sandy debrites. Turbidite-dominated depositional units display a heterogeneous distribution of facies and reservoir properties. Although the fields comprise more than one structure, they both have common oil–water contacts across the field area ( Figs 5, 6). In the Balder Field there is also a common gas–oil contact, whereas in the Jotun Field only the Tau structure has a gas cap. The latter, combined with minor differences in the level of the oil–water contact observed between individual wells in Tau, may indicate that there is some segmentation of the reservoir. This may impact reservoir performance during production with respect to sweep efficiency. Pressure plots from both fields support the field-wide contacts observed in wells ( Fig. 14). Both fields are in excellent communication with a large regional aquifer, as documented by the pressure decline over time ascribed to the production from the Early Tertiary Heimdal and Frigg gas fields ( Fig. 14).
Pressure plots: (a) Balder Field; (b) Jotun Field.
Reservoir parameters
Slides, slumps and associated sand injections are interpreted to be critical in establishing the connectivity conduits necessary to link the different mounds and other accumulations in the Balder Field ( Bergslien et al. 1996). A combination of sub-surface and outcrop data has been used to map sand injections in the Balder Field ( Fig. 15). Sand dyke and sill complexes are observed from wells and seismic data above all drilled mounds. These sand injections are particularly important as they connect the thin sands deposited above the main reservoirs and create an intricate network with good vertical connectivity ( Figs 15d and 15e). The combined volume of deposited thin sands and injection sands can be a significant contributor to the volumes in place as seen both in outcrop and in areas of the Balder Field, where the net:gross sand ratio of the injection complex interval is in the 15–20% range. In Jotun, the B-18 well in Elli South was drilled horizontally into a Zone 4 sand injection/slump complex above a major massive sand (Zone 3) with a thin oil column (5–10 m). The high water-cut in the well is interpreted to be a result of the good vertical connectivity partly due to sand injections. Fault juxtaposition of sands along slump faults may also contribute to the good vertical communication experienced across the shale separating the Zone 3 massive sand and the Zone 4 slump and injection sand complex.
Sand injections in outcrop and subsurface: (a, b) Puy in Provence, illustrating typical geometries of a sand injection complex; (c) W–E seismic section from Balder Field inversion cube (Mound 2 and 4); (d) schematic cross-section illustrating a network of injected and thin deposited sand with good connectivity and a common gas–oil contact; (e) UV light core photo highlighting strong oil saturation in injected sand above Mound 6 (well 25/11-19S).
Development Strategy and Production Performance
Balder and Jotun are both produced through floating production units (FPSOs). The Balder development consists of four sub-sea templates whereas Jotun utilized a fixed wellhead platform. The subsurface development strategy was chosen based on the interpreted reservoir geology and overall data quality at the time. Well density, horizontal lengths, positioning, trajectories and completion strategy were planned by an integrated team, with personnel from geoscience, reservoir engineering, sub-sea engineering, drilling and completion working together to optimize the development and maximize the capture and oil rates from the fields. For each well and development area within the fields, the technical risk and costs were evaluated versus benefits in terms of production and rates when considering the number and positioning of wells. The integrated work-team approach was a critical factor in the success of the development strategy designed to optimize the value of the asset, both from a resource recovery and an economic standpoint.
Balder well strategy
The major challenge in developing the Balder mounds is the relatively thin oil columns combined with the complex reservoir geometry. Reservoir simulation models and a five-month production test demonstrated that water coning would occur shortly after production start and would be quite severe in the massive sand mounds. The initial development strategy for the field was to drill a combination of vertical and horizontal producers. Pilot holes were planned to gather reservoir data and to narrow the uncertainty in the top reservoir definition. With these data, the completion sections were optimized and drilled with the maximum structural elevation above the oil–water contact. Attic oil loss was minimized and the timing and severity of water coning were delayed. In the planning of horizontal wells, there is a trade-off to be made between length in the reservoir versus stand-off from the oil–water contact ( Fig. 16). The interpreted top reservoir geometry within the mounds is therefore critical in deciding on optimal well trajectories. Sand control is required due to the unconsolidated nature of the Balder Field reservoirs. This was achieved through gravel packing vertical wells and wire wrap screens in horizontal wells. The initial horizontal wells are fairly short (400–800 m). The vertical producers are low angle wells (<45°) with ten-metre completions at the crest of mounds.
Balder Field well strategy. (a) D10 pilot and producer as planned from initial seismic interpretation; (b) inversion seismic data with planned D10 pilots and producer; (c) schematic cross-section illustrating results of the D10 wells; (d) D10 location map.
The initial development plan was to pre-drill several wells, tie them into the FPSO and gain some early data on production performance. Additional wells would then be drilled. With this strategy, early production history would be available for optimization of subsequent wells in the initial development. However, as a result of the late arrival of the production vessel, it turned out that all initial wells were drilled prior to production start-up. Without this early production history, the development strategy was evaluated and changed to include more horizontal and less vertical wells, based on analogue field performance and reservoir simulations. The initial development included 14 wells drilled from the four sub-sea field-sites; seven horizontal oil producers, three vertical oil producers, two vertical water injectors, one vertical gas injector and one vertical water-source well ( Fig. 2). The vertical producers were shut in early due to high water-cuts, and the two Mound 1 wells have thereafter been successfully re-drilled as horizontal producers.
Pilot holes were drilled at all production well locations with the exception of the A-6 well on Mound 1, which was drilled very close to the exploration well 25/11-8 and the long-term test production well 25/11-14SA. The pilot holes were drilled to accurately position the production well trajectories within the complex reservoir and to gather high quality log and core data. The latter was important as high quality core data were only present from Mound 1 by the Zone IB long-term production test well, 24/11-14S (1990) and the Mound 6 Zone II appraisal well 25/11-19 (1995). Log and core data from earlier exploration and appraisal wells are of lower quality than recent wells due to improved techniques in drilling (oil-based mud), log and core measurements, core preservation and positioning etc. One striking example of the difference in data quality through time is illustrated by the recognition of the Balder Field gas cap. The gas cap, now documented field-wide by MDT pressure plots, wireline and measurement-while-drilling logs, was not recognized in the field until the final appraisal well 25/11-19 was drilled in 1995. It detected gas both by classic neutron-density ‘crossover’ separation and by MDT sampling and pressures. Nearly all the previous 17 wells, drilled over a period of 26 years, had penetrated thin sands in the gas cap without detecting gas due to borehole washout in water-based mud and lack of logging resolution in thin beds. Thin bed reservoir characterization in the Balder gas cap is discussed in Haynes et al. (1999).
The development of the southern part of Mound 6 by the 25/11-D 10 wells is a good illustration of the Balder Field well strategy and performance. Two pilots were used to optimally position the producer as close as possible to the top of the reservoir sand. This was done to minimize water coning in the thin Zone II oil interval ( Fig. 16). The producer has a c. 800 m long horizontal section with 650 m of net sand behind the screens. The heel section penetrates and produces from a large-scale, low-angle sand injection. The stand-off from the oil–water contact is 10–39 m. Even if the stand-off along most of the well bore is low (10–15 m), the well is performing better than anticipated with a water-cut just above 50% at a rate of 16 000 BBL liquid per day.
A gas injector (D-1) was drilled into the gas cap over Mound 6 to store the associated gas produced with the oil. Water injectors were drilled in Mounds 2 and 4 as it was interpreted that they would have limited communication to the regional aquifer ( Fig. 2). A water-source well (A-8) was drilled into the Utsira Formation. This well has not been activated as the water-cut in the wells increased faster than predicted, producing large enough volumes for the two water injectors.
Jotun well strategy
At Jotun the appraisal drilling programme was designed to collect high quality stratigraphy, log and core data from all three structures. The appraisal strategy was a deliberate and critical element in the successful ‘fast track’ development of Jotun, maximizing data collection utilized for development planning. The 25/8-8 well was planned and drilled as a deviated well (8S) with two side tracks (8A, 8B) to appraise cost effectively the main part of the Tau structure and the eastward sand pinch-out in the area ( Figs 1, 3, 6 and 12). These well data, combined with the direct seismic mapping of the top reservoir and the less complex reservoir geometry, reduced the need for additional pilots in the field as compared to Balder. The B-2 well was drilled as a combined pilot for the B-1 and B-2A producers in the Elli structure. Another pilot was drilled in Tau (B-4) at the heel of the B-4A producer ( Fig. 3). In addition, innovative ‘toe pilots’ were drilled to improve definition of top reservoir and architecture in both Tau and Elli South ( Fig. 17a). The cost-effective ‘toe-pilots’ were drilled as extensions to horizontal wells in areas with uncertainty in terms of seismic, structural or stratigraphic definition. The development plan included only horizontal producers. The 12 initial oil producers were drilled based on 3D seismic interpretation integrating the pilot, exploration and appraisal well penetrations for reservoir and stratigraphic control. The 3D geological model and reservoir simulation was used in conjunction with well and seismic data to optimally locate and position well trajectories. The size and reservoir architecture of the Jotun structures resulted in the drilling of longer horizontal sections than the Balder mounds. The initial Jotun horizontal producers vary in length between 742 m and 1243 m of section, with net sand ranging from 149–798 m (net:gross of 0.2–0.9). All producers are horizontal open-hole completions with screens for sand control.
Jotun Field well strategy: (a) well 25/8-B7 producer with toe pilot; (b) cross-section from the 3D geological model along the planned B15 trajectory.
The reservoir geometry of the Elli structure is very similar to the mounds of the Balder Field, and thus a similar well strategy was applied. Elli production wells are drilled high on structure to maximize stand-off from the oil–water contact. This is because both horizontal and vertical connectivity in the debris flow-dominated deposits are very good due to the uniform reservoir architecture and distribution of reservoir properties (net:gross, vertical to horizontal permeability ratio etc.). The wells are all performing at high rates, 15 000 to 21 000 BBL liquid per day. Wells with 25 to 35 m of stand-off experienced water-cut within the first six months, whereas the B-2 well, located 42 m above the oil–water contact at the crest of the structure, did not produce water for 18 months. In addition to the greater distance from the oil–water contact, a shale baffle may also have helped to delay the flow of water to this well.
In Elli South the wells were also placed as high as possible, but the low relief structure limited the stand-off from the oil–water contact to just over 10 m ( Figs 6 and 10). Risk for early water coning was recognized but the shale separating the Zone 4 sands from the thick massive Zone 3 was interpreted to be a potential baffle. Special caution was exercised to avoid penetrating this shale. The two oil producers are completed in Zone 4 slump and sand injection complexes with a low stand-off from the oil–water contact (11–13 m). Unfortunately, the vertical connectivity turned out to be high and the wells coned water within the first months. 25/8-B11 well reached 50% water-cut after 11 months and B-18 three months after production start-up. The difference between the wells may reflect vertical connectivity or coning water resulting from nine months of B-11 production prior to B-18 start-up. The effective vertical connectivity across the shale interval is likely to be caused both by sand injections from the underlying Zone 3 sands and fault juxtaposition between the Zone 3 and Zone 4 sands.
The development strategy for Tau is challenging. It is a three-phase reservoir simulation combined with heterogeneous reservoir architecture with a fairly low vertical to horizontal permeability ratio and significant lateral variation in sand quality. Vertical flow barriers/baffles are expected as turbiditic depositional facies with laterally extensive shale dominate at Tau. The well trajectory strategy applied to Tau includes utilizing structure to cross as much stratigraphy as possible whilst drilling, in addition to optimizing the stand-off to the water below and the gas above the wellbore ( Figs 17a and 17b). Both the geological 3D model and the reservoir simulation model were important tools in designing well trajectories.
One water injector (B-10) was drilled between Elli and Tau, with the primary goal of supporting the Tau oil producers. The planned B-12 producer in the northern part of Tau found poorer sand quality than anticipated and was not completed as a producer. Instead it was converted into a water disposal well in the shallower Utsira Formation ( Fig. 3). As the early production data indicated the aquifer support was adequate, two of the three planned water injectors were omitted from the initial development phase. The strength of the regional aquifer was confirmed as the field was produced without any water injection for nearly two years with no observed dramatic pressure decline.
Conclusions
The Balder and Jotun fields both contain sediments transported eastward from the East Shetland Platform and re-deposited in the Viking Graben submarine fan complex as a response to the Tertiary tectonic activity associated with the opening of the Atlantic Ocean. The fields are located on the Utsira High close to the distal, eastern margin of the fan system. Reservoir sands originated from density gravity flows, and even though the fields are located close to the distal sand pinch-out, they generally display high net:gross ratios and contain thick high quality reservoir sands (>100 m). Both fields display compensational stacking reservoir architecture consisting of units with debritic and turbiditic origin separated by blanketing hemi-pelagic shales. The Balder Field Paleocene mounds and the Elli and Elli South structures of the Jotun Field are dominated by sandy debrites causing mounded reservoir geometry and a uniform internal distribution of reservoir properties. The younger, Early Eocene Balder Formation reservoir in the Balder Field and the Tau structure in the Jotun Field display heterogeneous reservoir properties with depositional facies partly dominated by turbiditic units, which generate a more tabular reservoir architecture.
In the Balder Field an intricate interaction between deep-marine deposition and soft sediment deformation processes generated a complex network of reservoir compartments. Slumping, sliding and sand remobilization with associated sand injections into overlying shale were the dominant deformation processes, occurring mainly during the early Eocene. The sandstones have excellent reservoir properties, share a common pressure system and field-wide fluid contacts. The connectivity throughout the field is to a large degree related to the sand remobilization process with associated sand injections establishing cross-stratal communication.
The Jotun Field displays greater lateral differences in reservoir quality and depositional patterns between the individual structures than those observed at the Balder Field. This may be related to a lower degree of post-depositional deformation and sand remobilization, but perhaps more importantly, is a result of a larger variation in primary depositional facies as channel complexes are recognized in Jotun. A large volume of the oil reservoir in Elli South exhibits a complex reservoir architecture due to sand remobilization. Occasional minor sand injections observed in Elli and Tau are likely to play only a minor role with respect to vertical reservoir connectivity.
The two fields generally demonstrate good to excellent lateral and vertical connectivity, both within and between individual structures. This is documented by field-wide fluid contacts, uniform response to regional aquifer pressure decline and production data. However, the gas cap at Tau is difficult to explain. Sand injections associated with larger-scale sand re-mobilization are likely to be an important factor affecting the connectivity within both fields. However, low permeability reservoir heterogeneity does exist, particularly within units dominated by thinly bedded turbiditic facies. These units have the potential to negatively affect fluid flow and sweep efficiency.
The reservoir quality and architecture governed by the facies recognition of the submarine fan depositional environment have impacted the subsurface development strategy with respect to well and completion type, number and placement of wells, length and geometry of trajectories. A multi-disciplinary team approach between geology, reservoir engineering, sub-sea engineering and drilling was utilized to address the subsurface development. This proved to be highly effective, both in the planning and the execution stages.
The key differences between the Jotun and Balder Field are:
the Balder Field is a complex cluster of small accumulations, whereas the Jotun Field has three larger structures;
the top reservoir is imaged as a reflector in the Jotun 3D seismic volume, whereas in the Balder Field top reservoir structure maps are constructed by isochoring from the depth-converted Base Balder seismic reflector;
a field-wide gas cap is present in the Balder Field, while in Jotun only the Tau structure has a gas cap;
the complex slump, slide and sand injection architecture seen across the Balder Field is only significant in the Elli South structure at Jotun;
the Balder Field shows vertical change from sandy debrites in the Heimdal Formation to turbidite-dominated reservoir facies in the Balder Formation. These depositional facies are seen to stack laterally within the Heimdal interval in the Jotun Field;
the fast track development of the Jotun Field benefited largely from knowledge generated through the long exploration and appraisal history of the Balder Field.
Just like two sides of a coin, the Jotun and Balder fields are similar, originating from the same depositional system but displaying important differences. They are linked together through the regional aquifer and by being part of the same submarine fan complex. However, there are also important differences, which have affected the development strategy concerning field history, number of structures, data quality, hydrocarbon composition, level of fluid contacts, complexity of reservoir architecture and properties.
Acknowledgments
I would like to thank ExxonMobil and the Jotun Field partners Enterprise, Conoco, Petoro and DNO for allowing me to publish this paper. The paper was first presented to the Exploration Branch of the Norwegian Petroleum Society (NPF), in Oslo, November 9th 2000.
Balder and Jotun – two sides of the same coin? A comparison of two Tertiary oil fields in the Norwegian North Sea
