The giant Albian upper Burgan reservoir in northern Kuwait is a major productive interval in both the Raudhatain and Sabiriyah fields. It is planned to increase production from the upper Burgan by over 300 percent through the application of field-wide waterflood. Phase 1 of seawater injection was started in 1999. A comprehensive geological study of the upper Burgan has included reviews of all core data, biostratigraphic information, and geochemical and petrographic analyses. The results indicate that the reservoir sandstones were laid down in a complex series of paralic environments ranging from shallow-marine to valley-fill channels. The resultant layering scheme was integrated with all available dynamic data to best define the reservoir architecture and flow units, and to construct a fine-grid 3-D geological model. The geological analysis has had a significant impact on reservoir development, influencing waterflood strategy, pattern orientation and expected performance. In the Raudhatain field, where reservoir connectivity is relatively high, a peripheral waterflood will be implemented (at least initially), whereas the complex structure and stratigraphy of the Sabiriyah field dictate a pattern flood.
The Kuwaiti oil fields of Raudhatain and Sabiriyah (Figure 1) are situated on the western flank and crest, respectively, of the N-trending Kuwait Arch. Both fields contain the three main productive formations of the lower Burgan Sandstone, upper Burgan Sandstone, and Mauddud Limestone. In addition, the Zubair Formation is a productive unit in the Raudhatain field.
In 1995, the Kuwait Oil Company (KOC) began a program of work aimed at significantly increasing production from the North Kuwait fields by 2005. The production target requires major facility and drilling expenditure along with significant subsurface studies to optimize the secondary development of several giant reservoirs. The upper Burgan is one of the four major reservoirs in the Raudhatain and Sabiriyah fields that are expected to contribute to the overall production target. Although the reservoir has been in production for about 40 years, the offtake so far has been only a small part of the potential ultimate recovery. This fact, together with the excellent reservoir qualities, highlights the secondary development opportunities that exist for the upper Burgan. With the development opportunity highlighted, an accurately defined geological framework was needed in order to optimize development and recovery.
The upper Burgan study was a means of better understanding the heterogeneity of the reservoir and its potential impact on reservoir performance. A wide range of data was obtained and analyzed. The existing data (production and pressure history covering more than 50 years; about 1,600 ft of core from 13 wells; and 167 well logs) were mostly in paper form and had to be converted digitally. New data consisted of 3-D seismic (300 sq km); petrographic, biostratigraphic and geochemical analyses; and image log analysis of four wells.
GENERAL GEOLOGY OF THE RAUDHATAIN AND SABIRIYAH FIELDS
The Burgan Formation (Figure 2) has been informally subdivided into the lower, middle and upper Burgan members, with the lower and upper being the most important oil-bearing units. The lower Burgan member includes a very thick section of high-quality reservoir sandstone and contains most of the reserves. The upper Burgan member, although smaller, is also considered as a giant reservoir interval.
The lower Burgan to Mauddud interval represents an overall transgressive sequence. The lower Burgan was deposited in dominantly non-marine conditions whereas the upper Burgan is interpreted as having been deposited in paralic to offshore environments and has a complex architecture of alternating shoreface and channel-dominated units (Davies et al., 1997). The rapid vertical changes are interpreted as reflecting short-term or high-frequency changes in relative sea level. Overlying the upper Burgan is the Mauddud Formation that was deposited under platform-carbonate conditions. In northern Kuwait, the seal to the Burgan consists of shales and siltstones at the base of the Mauddud (Goodall et al., 1996). Both the Raudhatain and Sabiriyah fields are faulted anticlines (Figure 3). The structures are defined by well control (more than 200 wells) and 3-D seismic data. The 3-D seismic survey was acquired in 1996 and covered approximately 300 sq km.
The Raudhatian structure (Figure 3) is a faulted anticlinal dome with 65 to 90 ft of topographical relief (Carman, 1996). The 3-D seismic has defined the major faulting in the northern part of Raudhatain as NW-trending, whereas in the southwestern part of the field the faults trend northeast. Fault throws are highly variable and range from less than 30 ft to as much as 150 ft. Most faults have throws of 50 ft or less. Structural dips vary across the field with the western side averaging 4° to 6° and the eastern side a gentler 2° to 4°. The upper Burgan reservoir in Raudhatain has approximately 600 ft of structural closure.
The Sabiriyah structure (Figure 3) is an elongated, N-trending, faulted anticline containing numerous grabens and associated relay ramps. It is considerably more complex than the Raudhatain anticline. The fault throws range from less than 30 ft to more than 160 ft. The structural dip varies from as much as 8° on the eastern flank to 3° to 4° on the western flank. Structural closure on the upper Burgan is approximately 350 ft. The original oil/water contact is unknown and is most likely to vary from one fault block to another. Due to the complex nature of the faults and the thinner upper Burgan productive layers in the Sabiriyah field, it is expected that faults play a major role in reservoir performance.
Previous studies had indicated that the faults are steep to near-vertical (Brennan, 1991; Carman, 1996). In both the Rauhatain and Sabiriyah fields, the newly acquired 3-D seismic has been integrated with the well data, conclusively showing that the fault dips range from 45° to 60°. The data was integrated using the Tangent Circle Mapping method of Tearpock and Bischkte (1995). Neither the sealing nature of the faults nor the impact of faulting on fluid flow is understood in either field.
The present-day principle horizontal stress field is interpreted as being generally orientated northeasterly in both the Raudhatain and Sabiriyah fields. The trend is based on 2 image logs and on 57 wells that had 4-arm caliper data available. This trend agrees with previous publications (Carman, 1996) that used a combination of borehole image logs, 4-arm caliper, and fracture data from outcrops along the Jal al Zor Ridge.
The methods used to define the sedimentology of the upper Burgan reservoir were: (1) detailed core descriptions; (2) integration of core sedimentology with conventional and image logs; (3) development of a sequence stratigraphic sedimentological model; and (4) development of a correlation framework that honors the sedimentological model and biostratigraphic data. Throughout the sedimentological study, the available dynamic data was continually relied upon to ensure that the overall product agreed with the historical reservoir performance. Dynamic data consisted primarily of:
Individual well-production results.
Pressure tests: 12 Pressure Build Up (PBU); 21 Shut-in Bottom Hole Pressure (SIBHP); 12 Repeat Formation Tester (RFT).
Cased-hole logs: 21 Thermal Decay Time (TDT); 6 Production Logging Tool (PLT).
In addition to the sedimentological study, the biostratigraphy of the reservoir was studied and an Iatroscan geochemical study defined the distribution of asphaltene percentage. Petrographic descriptions were also made and comprehensive image log and standard log analyses undertaken. The studies were made in parallel and the results integrated as appropriate.
Detailed core descriptions were the key source of data for the sedimentological study from which the following three schemes were devised:
A comprehensive lithotype scheme was developed based on the detailed core descriptions.
Following calibration of the core lithotypes with wireline log characteristics, a simpler log facies scheme was devised that took account of the lower vertical resolution of conventional wireline logs, and the non-unique log signature of many core-based lithotypes.
Additionally, an image facies scheme was established and this was combined with the detailed core lithotype scheme to provide a more-detailed picture of the four wells that have borehole image logs.
Within each of the three schemes, essential groupings of components define larger sedimentary bodies. The groupings facilitate the identification of a series of genetic elements that were interpreted as representing specific depositional settings with distinct depositional geometries. These are likely to define permeability architecture and, therefore, to strongly influence fluid movement during production. Environmental interpretations for the cored sections have been based on diagnostic sedimentary features, vertical trends and the vertical juxtaposition of facies types, taking full account of paleoenvironmental information contained within the palynological data (Jones, 1997). The interpretations have been extrapolated to uncored sections on the basis of interpreted log facies, vertical log trends (for example, cleaning-upward), sequence stratigraphic organization and vertical juxtaposition within the reservoir.
UPPER BURGAN RESERVOIR
The upper Burgan reservoir is approximately 150 ft thick and has a net-to-gross sand ratio ranging from 0.2 to 0.9 (Figure 4). Logs indicate several separate sand intervals that vary significantly in thickness and character from one well to another. The individual sand intervals vary in thickness from 5 to 80 ft. The consolidated sandstones are well sorted, very fine to fine-grained sublitharenites. Clay content is typically low (<20%) and cementation essentially non-existent. Reservoir quality is generally controlled by the content of the ductile component. The hydrocarbon-bearing intervals are between 7,400 and 8,000 ft deep and are normally pressured. The oil has an API value of 25° to 30° with a viscosity of 1 centipoise at the reservoir temperature of 170°F. The fluid properties vary significantly around the field and vertically within the reservoir. Pressure support to the reservoir is delivered by the surrounding aquifer. Aquifer support tends to be highest on the western flank of Raudhatain and steadily decreases to the east. In the Sabiriyah field, pressure support is considered poor on the eastern flank. Historically, the wells produced under natural flow conditions.
Productivity indices (PI) vary throughout the reservoir and typically range from 4 to 15 barrels of oil per day/pounds per square inch. In general, the higher PI wells are located in the crestal area whereas the flanks of the structure consistently have lower productivity. The decrease in productivity is mainly due to decreasing fluid quality and, to a lesser degree, to a decrease in reservoir quality. Historically, wells that produce from the upper Burgan reservoir are completed as duals with another reservoir unit. The completions vary in type from tubing long string to annular completions.
Production and Pressure History
Reservoir Production History
The upper Burgan reservoir is productive in both the Sabiriyah and Raudhatain fields. The fields were discovered in 1955 (Milton and Davies, 1965; Al-Rawi, 1981; Brennan, 1991; Carman, 1996). Production from the upper Burgan reservoir began in 1959 and 1970 in Raudhatain and Sabiriyah, respectively. In both fields, it represents a giant reservoir with greater than 1 billion barrels of original oil-in-place.
Historically, upper Burgan production from both fields shows variations in oil offtake rates, believed mainly to be associated with the offtake management strategies. Raudhatain upper Burgan production has been from 10,000 to 15,000 barrels of oil per day (bopd), with peak rates of about 30,000 bopd (Figure 5a). Production from the upper Burgan reservoir in Sabiriyah has averaged from 2,000 to 5,000 bopd with a peak production rate of nearly 15,000 bopd in mid 1995 (Figure 5b). In both fields, upper Burgan production has historically been characterized by periodic steep production declines.
Reservoir Pressure History
The initial reservoir pressure in the upper Burgan is estimated to have been 3,855 pounds per square inch gauge (psig) for both fields. Pressure data from PBU, SIBHP and RFT surveys indicate that pressure has declined substantially from this original value (Figure 5).
The limited spread in the pressure data for the upper Burgan reservoir suggests reasonable areal connectivity across the Raudhatain field (Figure 6). Aquifer pressure support is inferred by material balance and validated by recent Open Hole Logs and TDTs. The reservoir pressure is in decline and pressure support is clearly required to support any significant increase in the offtake rate. Formation pressure data suggest that shale/mudstone layers act as baffles to vertical pressure communication (Figure 7a). Pressure differences, from RFTs, of 20 to 30 psi are typical between separate sand bodies.
In Sabiriyah, there is a large range of reservoir pressure (Figure 6). High production rates during the post-Iraqi invasion period (1992 to present) have further contributed to the wide range in present-day pressures (Figure 5b). The large range suggests poor areal connectivity across the field, particularly in the crestal area where most offtake has occurred. Vertically, pressure differences of up to 500 psi (based on RFT data) are observed in the individual sand bodies (Figure 7b). The pressure variation is due to the complex relationship between perforation intervals, offtake and reservoir compartmentalization. Pressure maintenance will be necessary to support any significant increase in offtake.
The description of the reservoir is based on the detailed study of 13 original and newly acquired cores totaling about 1,600 ft. The cores were cleaned, slabbed (as necessary) and reboxed prior to description. Much attention was given in all phases of the curation to minimizing misplaced core and honoring previously missing sections.
The core descriptions recorded lithology, clay content, grain size, texture, composition, color, porosity, permeability and fracture occurrence on a 1:50 scale. The core descriptions provided the basis for the characterization of the reservoir’s lithotypes.
The lithotypes are the basic building blocks of the genetic elements. Together, they define the depositional architecture of the reservoir. Each lithotype has a distinctive textural fabric that can be related to reservoir quality and wireline log characteristics. Plate 1 illustrates 11 major lithotypes defined in the lower, middle and upper Burgan members (Davies et al., 1997).
Figure 8 summarizes the essential features of each lithotype and includes graphical representations identifying significant characteristics that can be compared to the detailed core descriptions. As an example of the lithotype scheme, Figure 8 and Plate 1 illustrate the cross-bedding (trough and planar) that defines the lithotype Ss(x). Figure 8 also depicts the presence of coarser-grained grainflow laminae that produce a significant degree of heterogeneity in the coarsest-grained, highest-quality sandstone of this lithotype. Note that this is much more common in the lower Burgan than in the upper Burgan.
The lithotypes provide evidence of the paleoenviroments. The graphical representation of parallel-laminated sandstones (Sl) indicates that Sl1 is coarser-grained than Sl2. This distinction is of environmental significance as lithotype Sl1 is almost exclusively associated with cross-bedded sandstones (Ss(x)) and is interpreted as a high-flow regime stream-flood facies, whereas the lamination in Sl2 is produced by preserved clay laminae. The Sl2 lithotype is an important, though relatively minor component of the shoreface sandstones genetic element. Similarly, the Sc lithotype is normally characterized by discontinuous bed-parallel laminations picked out by plant material and carbonaceous shale partings. An important variant is dominated by inclined, burrow-modified laminae.
The three varieties of bioturbated sandstones (Sb1, Sb2 and Sb3) provide important environmental clues. The vertical burrows that dominate Sb1 indicate high-energy, shallow-water conditions whereas the more diverse assemblages with Asterosoma (Sb3) represent quieter, generally more offshore settings. These general interpretations are supported by the overall depositional context as well as palynological data. However, it should be noted that burrows such as Asterosoma are not exclusively marine forms, and can be an important component of estuarine/marginal marine environments. Of environmental significance are burrowed clay drapes on foresets of cross-stratified sandstone (Ss(x)) and the heterolithics (H) that provide evidence of an important tidal influence, inclined lamination (particularly in heterolithics and mudstones) that provide evidence of lateral accretion, and rootlets that indicate plant colonization and are almost the only evidence for paleosols in the Burgan Formation.
Log Facies Scheme
In order to extend the interpretations of lithotypes beyond the sections that contain core, a relationship between wireline response and lithotype was established. Log facies were interpreted in 50 key wells across both fields in addition to the 29 cored wells. Conventional wireline logs do not have sufficient resolution to identify many of the lithotypes observed in the core. For this reason, a simpler log facies scheme has been devised, though this is based on the original core lithotypes. The essential features of the log facies scheme are presented in Table 1. In particular, it identifies the equivalent core lithotypes to each log facies, and to which genetic element they are likely to belong (see Genetic Elements below). The most important log facies are Sx, Sc, Sb and M.
The primary log used in the log facies scheme is the gamma-ray log. However, evidence was assessed from all logs, especially the neutron-density combination where available, and the SP, sonic, resistivity and caliper.
Image Facies Scheme
One of the important elements of this study was a borehole image analysis of four wells (Taylor and Tribe, 1997). The image facies displayed in Table 2 were based to a large measure on the core lithotypes, although only one of the four wells was extensively cored through the upper Burgan reservoir. This Table summarizes the image facies scheme, outlines the image fabrics, and indicates both equivalent core lithotypes and the genetic elements to which they are likely to belong. The borehole image study also yielded valuable paleocurrent data for the channel sandbodies.
Both cored and uncored sections have been divided into a series of genetic elements (or depositional packages) interpreted as representing specific depositional settings (Table 1). Eleven genetic elements were identified in the whole Burgan Formation although some are of minimal importance. Table 1 shows the seven most important genetic elements and log facies in the upper Burgan. Associations of genetic elements are more important for identifying gross depositional environments (for example, floodplains and deltas), rather than individual elements, such as channels, that are found in many different settings. They form the basis for the depositional models presented in this study.
Each genetic element has a distinct depositional geometry. The geometries are important in that they define large-scale (for example, sandbody-scale) permeability architecture. Hence, they can be expected to have a fundamental influence on fluid movement during the field’s production life. Knowledge of their geometries and dimensions, combined with well spacing and log signature, is a key input to reservoir modeling. Figures 9a and 9b illustrate the genetic elements associated with channel sandbodies and estuarine/marginal marine mudstones and sandbodies.
The channel sandbodies, which include major and minor marine-influenced and laterally accreting muddy channels, are typically characterized by sharply based units and a consistent grain size, and have a thickness range of about 10 to 50 ft. Major channels are more than 20 ft thick. The dominant lithotypes present are Sx and Sc. Many sand-filled channels fine upward into burrowed or carbonaceous mudstones. The channels have a general northeasterly trend based on the measurement of paleoflow directions from detailed image log analysis and overall net-sand trends. In many wells with high net to gross values (>0.7), the individual channels have amalgamated to form areas of nested channel-fills. In such areas, vertical communication is enhanced whereas the degree of lateral communication remains unknown.
Individual major channels present in one well are commonly very difficult to trace between the 79 key wells (cored and uncored). This suggests that the individual channels are typically less than 1,000 m wide and may be highly sinuous. Correlation is more difficult in the Sabiriyah field than in Raudhatain.
Estuarine/marginal marine mudstones and associated sandbodies have a high level of heterogeneity and are commonly associated with minor marine-influenced sandstones. These genetic elements commonly show evidence of regular tidal activity and have abundant carbonaceous material and amber. The sandbodies are less than 10 ft thick and are likely to be highly discontinuous and poorly connected.
Changes in Sea Level
The wide variety of genetic elements and their complex vertical organization means that no single depositional model is representative of the Burgan Formation, though it can be classified as paralic (deltaic to shallow marine). In this study, a suite of depositional models is presented with each representing the response of the depositional system to changes in relative sea level (Figure 10). Several different orders of relative sea-level change influenced deposition. An overall long-term (low frequency) rise in sea level occurred. Its net effect was that fluvial processes were dominant in the lower part of the lower Burgan member but that the influence of marine processes became stronger upward. Medium-term changes in relative sea level are shown by the advance and retreat of the upper Burgan depositional system. High-frequency fluctuations in relative sea level gave rise to the complex reservoir architecture of the upper Burgan member. The sequence of depositional models considers this vertical evolution with the positioning of the models on Figure 10 intended to be a reflection of relative sea level.
During lower to middle Albian times, the Raudhatain/Sabiriyah area was an offshore environment. The paleostrand line is interpreted as having a general northwesterly orientation to the southwest of the two fields. The middle Burgan member was deposited at this time and is dominated by a thick section (140–160 ft) of stacked progradational muddy shorefaces. During the middle to upper Albian, this muddy shoreface environment evolved into the complex upper Burgan member. Core and wireline log correlations indicate that the upper Burgan was deposited in a complicated alternation of depositional environments that included major marine-influenced channel and estuarine settings, and occasional periods of relatively high sea level as recorded by thin marine mudstones and/or shoreface sandstones.
Although there is positive evidence of estuarine processes, the more regressive, hence more constructional, lower part of the upper Burgan may have been partially deposited by tidally influenced deltas (Figure 10-F). Estuarine conditions (Figure 10-C) would have been more prevalent during the retreat of the entire depositional system, and so would become more dominant in the uppermost part of the upper Burgan member. There is abundant evidence of channel sandbodies cut into marine mudstone in the uppermost part of the upper Burgan (Figure 10-G) and micropaleontological data suggest that some of the enclosing marine mudstones were deposited in water depths of a few tens of meters (Jones, 1997). This provides good evidence of significant drops in relative sea level that facilitated channel incision. The channel-fills at the top of the upper Burgan are true, although small, incised valley-fill deposits. The major channels are believed to have northeasterly trends based on paleocurrent directions obtained from image log analyses. As relative sea level continued to rise overall, marked retrogradation occurred (Figure 10-D). The final extinction of the clastic-dominated depositional system during a continued rise in relative sea level resulted in the establishment of carbonate deposition and the accumulation of the carbonate-rich Mauddud Formation (Figure 10-H).
Quantitative data from core, thin-section, scanning electron microscope (SEM) and X-ray diffraction (XRD) analyses were used to determine the textural and mineralogical composition of the upper Burgan reservoir and establish the principal geological controls on reservoir quality. Twenty-three samples from three wells were analyzed. The samples were taken from all of the major lithotypes in the reservoir.
Petrographic analysis of the cored upper Burgan intervals shows that the reservoir consists of mineralogically mature (that is, quartz-rich), very fine- to medium-grained sandstones and siltstones in which the ductile content is highly variable (0–70%). The sandstones are quartz-rich (typically >70%) and feldspar poor (typically <5%) and are classified as predominantly sublithic/subfeldspathic to quartzose. Other framework components include polycrystalline quartz, extrabasinal rock fragments, heavy minerals, amber and very rare bioclasts.
Detrital clay is the principal ductile phase and typically comprises 12 percent of the total rock volume. Its presence is dependent on lithotype (Table 3). In the cleaner lithotypes (Ss(x), Ss, Sm and Sc), the detrital clay content is commonly less than one percent. In contrast, in the more bioturbated lithotypes, detrital clay may make up as much as 54 percent of the total volume.
Illite is present as a red/brown organic-rich clay commonly found in the more bioturbated sandstones/heterolithics and mudstones. It is present in both grain-rimming and pore-filling settings. In addition, green microporous clay of mixed composition (illite/kaolinite/chlorite/smectite, as validated by XRD and SEM) has a pore-lining to pore-filling role.
Organic matter, glauconite, mica (principally biotite and chlorite) and ductile rock fragments (mudclasts, smectitic grains and rare peloids) are additional ductile components. Glauconite is particularly abundant in some upper Burgan sandstones.
Overall, the diagenetic modification of the upper Burgan reservoir is minor, although local cementation by siderite, calcite, ferroan dolomite and authigenic quartz has occurred. Of these cements, siderite is the most obvious in core and on wireline logs. It is primarily replacive of detrital clay, but is also present as a pervasive, microcrystalline pore lining and pore-throat obstructing cement. It is most common in (but not restricted to) the bioturbated and glauconitic lithotypes. The apparent preferential development of other carbonate cements within some flank wells highlights the possibility of differential (water-leg) diagenesis within these fields. However, this hypothesis must be tested by more detailed petrographical evaluation.
Because of the overall clay-rich nature and very fine- to fine-grain size of the upper Burgan reservoir sandstones, the ductile content is the fundamental control on reservoir quality (Figure 11a). Hence, permeabilities are a function of lithotype (Table 3). However, grain size is also critical to permeability, particularly within the cleaner and better-quality lithotypes (Figure 11b). Diagenetic modification is of only localized importance. Figure 11c is a cross-plot of permeability against porosity for the reservoir and shows the correlation of the cleaner sandstones with enhanced reservoir quality. Arithmetic core-permeability values from core-analysis data range from about 10 to more than 1,500 md (average 513 md). Core helium porosity values of the reservoir-quality lithotypes range from about 5 to 30 percent, with an arithmetic mean of 20 percent.
A quantitative palynostratigraphic analysis was made of 150 samples from the upper Burgan reservoir (Jones, 1997). The samples were chosen taking into account both the vertical and areal distribution in the fields. Vertical sampling was at approximately 30-foot intervals whereas the field distribution was constrained by the availability of core data. The objective of the biostratigraphic work was to help constrain the sequence stratigraphic interpretation from the sedimentological study and to provide biostratigraphic evidence for paleoenvironmental interpretation. This analysis was made in parallel with the sedimentological study.
Results from the palynological study revealed five correlatable biostratigraphic events in the upper Burgan reservoir to which chronstratigraphic as well as correlative significance can be attached. No one well has evidence of all five events but Well RA-D in Figure 12 shows four of them. In conjunction with dated overlying and underlying units, the stratigraphic relationships indicate that the upper Burgan is of middle to late Albian age (see Figure 2).
The palynological features that characterize the five events are as follows:
Bio-event 1: one of more of the following—base common Spiniferites multibrevis and common S. ramosus; base Xiphoridium alatum.
Bio-event 2: common occurrences of Afropollis jardinus, Araucatiuacites australis, Inaperturopollenites limbatus, and Oodnadattia tuberculata.
Bio-event 3: common leiospheres.
Bio-event 4: top Densoisporites velatus, Ephedrites barghorrni, Reyrea polymorpha and Seripea naviformis, and top common Oligosphaeridium albertense.
Bio-event 5: top Stellapollis barhoorni and Sestroporites pseudoalveolatus, top abundant Subtiltsphaera and base common leiospheres.
The bio-events provide a valuable check on the validity of the sequence stratigraphic correlation scheme. The biostratigraphical analysis is in excellent agreement with the correlation scheme defined in this study, and several of the correlatable markers appear to have diagnostic palynological signatures (Jones, 1997).
The biostratigraphic data alone suggest that the upper Burgan reservoir is a highstand systems tract characterized by regressive sediments and proximal palynomorph assemblages (Jones, 1997). The palynological data is particularly significant in that it identified a marginal marine-influenced signature in most of the samples analyzed.
The sedimentological and petrographical evaluations clearly demonstrate that primary depositional characteristics controlled the reservoir properties of the upper Burgan reservoir. Therefore, it follows that the reservoir layering would be controlled by the stratigraphy of the upper Burgan member, which essentially determines the large-scale permeability architecture.
The aim of the sequence stratigraphic evaluation was to build a well-correlation framework that could be extended to all wells in both fields. The well data set consisted of 167 wells (101 in Raudhatain and 66 in Sabiriyah). The basis of the reservoir layering is the identification of key correlatable surfaces that mark significant landward or basinward shifts in the large-scale depositional systems. The key surfaces can be confidently identified in core, have a biostratigraphic significance and many have distinct wireline log characteristics. Figure 12 is a correlation panel across the Raudhatain field.
Many candidate surfaces were identified using the extensive core coverage throughout the Burgan Formation. This process was extremely important as the sedimentological and palynological data indicate that many shale sections within key parts of the upper Burgan reservoir are of marginal-marine (specifically estuarine) origin. It also demonstrated that channel incision had eroded marine mudstones in many places. Hence, there are inherent risks in correlating estuarine shales that are likely to be discontinuous, and even some of the key marine shales are commonly missing through erosion and could not be identified in all wells. The surfaces used for regional correlation are a subset of the candidate surfaces, and are based on well-to-well correlation across both fields.
Many rapid facies changes were identified by the interpretation of log facies in the uncored sections of cored wells and throughout an additional 50 key wells. Naturally, facies changes in uncored sections are less certain. Marked changes in log facies were calibrated as far as possible with neighboring cored wells, and field-wide correlation proceeded on this basis. The wireline signatures, such as inflections, of correlatable surfaces in cored wells are the basis for identificating equivalent surfaces in surrounding wells.
Three main types of stratigraphic surfaces are present in the Burgan Formation.
Flooding surfaces: surfaces across which there is evidence of an abrupt increase in water depth; they form during periods of relative rise in sea level. These are assigned names according to the member and a decimal scheme (e.g. MB10, UB30).
Sequence boundaries: erosive unconformities or their correlative conformities formed during a relative fall in sea level. These have names that identify their position in the reservoir; for example, on a chronostratigraphic diagram UB25SB lies between the UB30 and UB20 flooding surfaces through it may erode down into older layers.
Transgressive surfaces: the lowermost flooding surfaces formed during relative rises in sea level; they cap an older topset section. Transgressive surfaces of erosion (TSE) are transgressive surfaces or flooding surfaces at which material has been removed by marine erosion. This erosion may have removed evidence of subaerial environments such as interfluves.
The sequence stratigraphy of the middle and upper Burgan is shown in Figure 13 that documents the changes in relative sea level and illustrates where missing sections occur due to exposure and erosion. It also summarizes the interpreted high-and low-frequency organization, presented in terms of depositional systems tracts.
The middle Burgan member can be assigned to highstand systems tracts. Most of the succeeding upper Burgan member is interpreted as a complex succession of high-frequency lowstand and highstand systems tracts that form the upper part of a low-frequency highstand systems tract.
The ultimate retreat of the clastic depositional system and the establishment of carbonate deposition are the product of a low-frequency transgressive systems tract. After the initial rapid advance of the clastic Burgan depositional system, its subsequent, punctuated retreat matches the interpreted longterm rise in sea level through the Aptian and Albian stages (Haq et al., 1988).
For the reservoir-layering scheme, zonal layers have been named according to the stratigraphic surface, flooding surface or sequence boundary, that defined its top. Because of the high-resolution layering in the upper Burgan, there are many places where picks for flooding surfaces and sequence boundaries are coincident and one or more zones are missing (Figure 12). This is due either to erosion by channel incision or where the zonal equivalent of a channel/valley fill is an interfluve surface. Some channels were deep enough to cut out more than one zone, particularly in the uppermost part of the section where zones are relatively thin. In these places, tabulated thickness data clearly indicate those zones that are interpreted as being present and those that are absent.
The layering scheme was constructed using sequence stratigraphic principles and constrained by the biostratigraphic analysis, the use of the available dynamic data, and paleoflow information deduced from borehole image analysis (Taylor and Tribe, 1997). The dynamic data consisted primarily of Repeat Formation Tester (RFT) pressures, Pressure Build Ups (PBU), and Thermal Decay Times (TDT).
The RFT data points are particularly valuable in constraining a layering scheme. RFT pressures on either side of a shale interval can give an indication of the extent of the shale. This observation is extremely valuable when combined with the available sedimentological and biostratigraphic data. Aquifer influx distribution, as noted by TDTs, is also helpful in defining the flow units. For example, the correlation of water influx in one sandbody with that in another sandbody in an offset well can help constrain sequence stratigraphic correlations.
Isopach and facies maps have been constructed for each zone. Isopach maps for the zones dominated by marine rocks (zones UB65, UB55, UB45, UB35, UB25) are a guide to both their areal distribution and the effectiveness of these intervals as vertical transmissibility barriers. Where one or more of these zones are missing due to erosion, there is likely to be much stronger vertical communication.
An example of a facies map (Figure 14) shows Zones UB70 and UB65 and, therefore is representative of the transgressive systems tract at the top of the upper Burgan member. It illustrates the complex distribution of the productive facies in these two zones. As the width of channels is typically less than 1 km, whereas the separation between key wells is typically greater than 1 km, individual channels (or valleys) cannot be mapped between wells using the present database. Channel systems are shown instead.
A total of 380 samples from 14 wells in the Raudhatain and Sabiriyah fields were analyzed by Iatroscan and the data used to define areal and vertical asphaltene trends in the upper Burgan reservoir in both fields. Iatroscan is an accurate yet simple geochemical technique used to investigate fluid type in terms of vertical and areal trends. The technique involves separating the core extract into four fractions: saturates; aromatics; resin-A (polars); and resin-B (asphaltenes) (Barwise, 1997).
Several important insights into fluid characterization and vertical connectivity resulting from this work are illustrated in Figure 15. These are:
Iatroscan analysis showed that there is no evidence of tarmat zones in the upper Burgan reservoir in either the Raudhatain or Sabiriyah fields. Tarmat zones are typically characterized by an asphaltene content of more than 80 percent.
The upper Burgan reservoir in the Raudhatian field is gravity segregated. The asphaltene content increased with increasing depth but it did not reach the 80 percent level indicative of tar mats. The gravity-segregated nature suggests that the reservoir has a large degree of vertical connectivity in order to allow vertical fluid communication over geological time.
The Sabiriyah upper Burgan reservoir does not appear to be gravity segregated and it is likely to have extensive vertical barriers to fluid communication. Clearly this observation matches the geological model.
3-D GEOLOGICAL MODEL
From the large amount of new data available, a fine-grid geological model was constructed for the Sabiriyah upper Burgan reservoir that integrated 3-D seismic, sedimentological, and petrophysical data. The faulted 3-D model was built with three benefits in mind.
Quality Control of the input data: a 3-D model is an excellent tool with which to check the quality of the structural surfaces, layer picks and petrophysical property integrity.
Visualization of the geology: in a multidisciplinary environment, a consistent picture of the reservoir adds significant value. 3-D visualization of the structure and reservoir-quality trends is invaluable in optimizing well locations and predicting water movement to aid in wellwork opportunities.
Simulation Input: with the construction of a fine-grid geological model, an updated reservoir description can be used as input for a simulator.
The 3-D model was constrained structurally and stratigraphically by the newly acquired data. Structurally, the constraint was the 3-D seismic over the entire field area. The major faults that have a displacement of more that 50 ft were included in the model to better assess the juxataposition of productive layers. So far, 3-D seismic attributes have not been used to constrain correlations due to the present inability of the seismic to see the internal reservoir architecture. Stratigraphic constraints resulted from the new layering scheme and the directional information from the borehole image analysis. The new layering scheme, described previously, was extended to 167 wells and incorporated all sedimentological and dynamic information. The construction of layer grids created the 3-D stratigraphic framework of the model. Within each layer, further sublayers were defined based on the inferred dominant depostional process. For example, in a layer that is dominated by amalgamated or major marine-influenced channels, a layering scheme of incision was invoked.
Populating the 3-D framework was based on deterministic interpolation of petrophysical attributes. The interpolation was constrained by the interpreted dominant paleoflow direction for each layer. Numerous petrophysical attributes, such as, porosity, net-to-gross sand, water saturation and permeability can be displayed. Stochastic modeling was not undertaken at this time but may be in the future.
DEVELOPMENT PLAN CONSIDERATIONS
The reservoir description has many implications for the future development of the upper Burgan reservoir in the Raudhatain and Sabiriyah fields. They include the identification of permeability barriers or baffles, sandstone connectivity and distribution, reservoir quality, high-permeability conduits, and net-pay identification. The main features of reservoir layering and internal heterogeneities likely to influence fluid movement through the Burgan reservoir are summarized in Figure 16. This figure contains information concerning the relative scales of heterogeneities and their degree of importance.
The analysis of reservoir geology identified several types of mudstone that have different depositional geometries. The obvious consequence is that different types of mudstones will be effective barriers at very different scales throughout the Burgan Formation (Figure 16). The complex pattern illustrated by RFT data is consistent with the depositional architecture, in that there are marked variations in the magnitude of pressure gradient shifts across individual mudstones. The main points are:
The major field-wide permeability barriers are marine mudstones overlying correlatable flooding surfaces (Figures 12 and 16). They are of less significance (though still important) in the upper Burgan member where many of them have been breached by downcutting channels.
Siderite-replaced glauconitic sandstones (and/or ironstones) form permeability barriers within the upper Burgan member, but are probably effective over the area of only a few wells due to common erosion by downcutting channels (Figures 12 and 16).
Intervals of estuarine mudstones undoubtedly form localized transmissibility barriers but they are difficult to correlate between wells (Figure 12) as most of them have preserved widths of less than 1 km, which is smaller than the well spacing. These dimensions reflect original depositional limits within drowned river courses implying that they are elongated along paleoflow directions and must be the product of channel erosion. It is likely that there has been a degree of amalgamation to produce some packages of estuarine mudstone with greater lateral dimensions (1–5 km). These are likely to form effective permeability barriers but difficulties of precise shale-to-shale log correlation mean that production/surveillance data are required to identify them. The present evidence is that incursions of marine mudstones overlying correlatable flooding surfaces are more likely to form laterally extensive transmissibility barriers than are estuarine mudstones. It should be expected that channel sandbodies separated by estuarine mudstones show lower pressure gradient shifts from RFTs than those separated by a major marine mudstone.
Smaller-scale baffles that are effective at the sub-kilometer scale are likely to be formed by mud-filled estuarine channels and the muddy caps to marine-influenced channels.
Mud-draped lateral accretion surfaces are likely to form baffles tens to hundreds of meters long.
Many small-scale heterogeneities seen in core and thin section are also likely to impair fluid movement. These include mud drapes on cross-bed foresets and the abundant mud that has been redistributed in bioturbated sandstones.
The geological analysis of the reservoir identified several different types of sandstone, each having its own depositional geometry. The main points are:
Major marine-influenced channels are the most important reservoir facies in the upper Burgan member. Individual channels can only be identified in single key wells, but are amalgamated to form significant sandbodies (>80 ft thick) that extend between several wells (Figure 12). Such amalgamated channel sections are likely to have highly labyrinthine reservoir architecture and their axes to be preferred water-breakthrough conduits. Presently available data in the form of paleocurrents derived from borehole image logs (Taylor and Tribe, 1997) and maps of dominant depositional elements (for example, Figure 14), suggest that paleoflow was predominantly toward the north and east.
Major marine-influenced channels are likely to provide important vertical high-permeability conduits to fluid movement where they breach low-permeability layers (Figure 16). In addition, they could cause major difficulties in predicting water breakthrough during future production. At a much finer scale, coarse-grained grainflow laminae represent high-permeability streaks that collectively could significantly influence water movement.
Development Plan (Peripheral versus Pattern Waterflood)
The development plan for the upper Burgan reservoir in each field was designed primarily on an understanding of the stratigraphic and structural complexity. In addition, productivity/injectivity estimates and other reservoir engineering data were considered (Abdullah et. al., 1998). The reservoir description was integrated with the available dynamic data to influence the development type and flexibility.
The initial upper Burgan development is to be by peripheral waterflood. The geological analysis indicated higher levels of connectivity within the reservoir (when compared to Sabiriyah upper Burgan) due to the presence of more amalgamated channels resulting in a higher net-to-gross interval. The better connectivity was further validated through integration with the 3-D seismic, RFT/PBU pressure data, and water production history. New well locations are being positioned and linked with other reservoir developments so that an easy change from the peripheral flood to an inverted 9-spot-pattern development is possible if reservoir performance under peripheral flood is less than expected.
The development of the upper Burgan is planned as an inverted 9-spot-pattern waterflood. The geological analysis indicated that a peripheral flood was not possible as an extreme variability in net sand existed between closely spaced wells (<1 km in places) in many parts of the field. This observation, combined with complex 3-D seismic structural interpretations ruled out the possibility of a peripheral flood. The observed geological complexity agrees well with the RFT/PBU pressure information and the production trends described earlier.
The pattern orientation for the upper Burgan reservoir in each field was influenced by interpreted paleocurrent direction and the Maximum Horizontal Stress (MHS). Borehole image analysis indicated the paleocurrent direction to be northeasterly. In addition, the MHS was estimated to be NE-trending based on Image Log analysis and borehole breakout studies. The patterns have been orientated such that the greatest distance between injector and producer is in the northeasterly direction. The increased distance may reduce the risk associated with earlier than expected water breakthrough.
Reservoir performance and recovery in paralic environments is a function of well spacing and net-to-gross value. It should also be noted that performance is typically over-predicted due to non-recognition of reservoir complexities in early planning (Dake, 1994). The geological analysis for the upper Burgan was considered fully in the performance prediction studies conducted within the Kuwait Oil Company.
Reservoir performance was predicted using the classic analytical technique of Stiles analysis (Craig, 1993) for fractional flow and vertical sweep. All available core data were used to assess the vertical permeability variation and the sedimentological trends that influenced the arrangement of the data into layers. The areal sweep was estimated by means of the technique described in Craig (1993) for theoretical values. These values were then discounted by a ‘compartmentalization factor’ that represents an educated estimate of the level of heterogeneity found in the reservoir. In the Raudhatain field, where reservoir connectivity is high, a 90 percent factor was used whereas in Sabiriyah, where reservoir connectivity is expected to be lower, a 75 percent factor was applied.
The integration of the recognized geological complexities with the predicted performance tools affected the predicted water-cut development and estimated recovery factor. The upper Burgan oil- and water-cut profiles and predicted recovery factor compare well with actual analog data.
The Kuwait Oil Company has embarked on a strategy to significantly increase oil production from the northern Kuwait fields. As part of the area development program, current production from the Albian upper Burgan reservoir in the Raudhatain and Sabiriyah fields is expected to be increased by more than four times. This dramatic increase in production will be achieved through the application of field-wide waterflood development. The upper Burgan reservoir is a complex paralic depositional environment whose understanding has required a comprehensive analysis.
A comprehensive geological study consisting of core description, petrographic and image log analyses, 3-D seismic, biostratigraphy, geochemistry and sequence stratigraphy was integrated with all available dynamic data to best define the reservoir architecture and flow units. The detailed description of the reservoir has resulted in an assessment of reservoir heterogeneities that are critical to the understanding of past fluid flow and future predictions. The traditional products of a layering scheme, plus 3-D seismic and petrophysical data, were then used to construct a fine-grid geological model based on the new description.
The geological study has had a major impact on the reservoir development strategy by influencing the development plan. Reservoir description influenced the choice of a separate development type for each field. In Raudhatain, the reservoir connectivity is considered high and a peripheral waterflood will be implemented, at least initially. In Sabiriyah, the complex stratigraphy and structure dictates a pattern flood. Pattern orientation was optimized based on the interpreted paleocurrent direction and the maximum horizontal-stress direction. In addition, all available core data and layering was considered in defining expected reservoir performance. Refined definitions of expected water-cut development and areal sweep were achieved resulting in excellent comparisons to actual analog performance data.
The revised reservoir description provides the development team with an improved understanding, an enhanced predictive capability, and the ability to optimize new well locations.
The authors would like to thank Dorothy Payne, Katy Taylor, Geoff Freer, Bob Jones, Allywn Siqueira, Chris Reddick, Hosnia Hashim, Nigel Rothwell and Martin Smith for their contributions to this paper. The comments and suggestions of the reviewers and editors are acknowledged. The design and drafting of the final graphics was by Gulf PetroLink.
ABOUT THE AUTHORS
Ahmed J. Al-Eidan is a geologist with the Kuwait Oil Company (KOC). He received a BSc in Geology from Kuwait University in 1992 and joined KOC in January 1993. For three years he worked as a Wellsite Geologist and then as a Development Geologist until 1999. He has been a member of the Prospect Evaluation Team since 1996. Ahmed is currently working as an Exploration Geologist and leader of the Exploratory Drilling Team.
W. Barry Wethington received a BSc in Geology in 1985 and an MSc in Energy and Mineral Resources from the University of Texas at Austin in 1989. His major professional interests are the applications of development geology to field development and reservoir management. He has worked in many mature fields on Alaska’s North Slope and elsewhere in the world. More recently, Barry has led a multidisciplinary team that focused on the development plans of the upper Burgan reservoir.
Roger B. Davies has 20 years experience in the oil industry. He has been involved in the investigation of clastic and carbonate reservoirs in many parts of the world, including the North Sea, Alaska, and the Middle East. He was employed by British Petroleum as a Sedimentologist from 1980 to 1994. Since then, Roger has worked as a Consultant and is an Associate of Badley, Ashton and Associates and the Energy Resource Alliance. He has a PhD from the University of Southampton. Roger is a co-author of GeoArabia Special Publication 2, Arabian Plate Sequence Stratigraphy.