Uganda's Albert graben due first serious exploration test

June 10, 2002
There's a growing sense of anticipation in the heart of Africa as Canadian independent Heritage Oil & Gas Corp. and partner Energy Africa Ltd. prepare to drill their first exploration well in the Albert graben of western Uganda.

There's a growing sense of anticipation in the heart of Africa as Canadian independent Heritage Oil & Gas Corp. and partner Energy Africa Ltd. prepare to drill their first exploration well in the Albert graben of western Uganda.

The Albert graben has been a focus of interest for petroleum prospectors for nearly a century, but previous licensees did not progress beyond geological and geophysical field studies, discouraged by the challenges of working in this area. It's only in the past few years that Heritage has overcome these challenges, thereby allowing a full exploration work program to proceed to the drilling of an exploration well.

New basin exploration wells are rare these days, but this previously undrilled and very attractive hydrocarbon province will now at last see its first significant exploration drilling campaign later this year.

Exploration locale

The Albert graben occupies the northern part of the western arm of the East African Ring the basin margin. Harris et al.2 described the seeps in more detail in 1956.

Click here to enlarge image

One major oil seep lies on Block 3 at Kibuku (Fig. 1). In common with all the seeps, the seep at Kibuku in the foothills of the Rwenzori Mountains is located at the contact of the sedimentary basin fill and the underlying basement.

Although the presence of hydrocarbons in the system has been known for some time, only one well, Butiaba Waki-1, drilled by African & European Investment Co. in 1937, has been drilled in the Albert graben.

Click here to enlarge image

Butiaba Waki-1 is close to the oil seep at Kibiro, on the eastern shore of Lake Albert, north of Block 3. The well drilled through the basin-bounding fault into basement rocks at 1,222 m and therefore cannot be regarded as a valid test of the basin. No logs were recorded apart from a lithology log, but oil shows were encountered while drilling.

Click here to enlarge image

Upper Tertiary rocks crop out in the south of Block 3, north of the Rwenzori foothills. Pickford et al.,3 mapped and dated these outcrops, establishing the basic stratigraphic framework for the area. Petroleum Exploration and Production Department of Uganda (PEPD) further refined this framework in various unpublished reports from 1996.

Click here to enlarge image

Since 1990 PEPD has acquired gravity and magnetic profiles in a series of traverses across the basin. Modeling of these data suggests a thickness of sedimentary section varying from 4 to 6 km in the deeper parts of the basin. Gravity profiles across Block 3 also show a major change of sedimentary thickness across an intrabasin fault, which suggests that structures may be present in the section.

Field work

The section at outcrop ranges from Middle Miocene, based on radiometric dating (11.5 MY), to Recent, summarized in Fig. 2.3 4 This modified stratigraphic column shows the addition of one new formation, Kasande, which was not mapped by previous workers.

The oldest section at outcrop is the Kisegi formation, a predominantly sandy unit, which is made up ft Valley (Fig. 1).

Block 3 lies in western Uganda, abutting the border with the Democratic Republic of Congo (DRC). The license covers:

  • The southern part of the Ugandan portion of Lake Albert;
  • A landward area, known as the Semliki flats, south of the lake comprising the eastern part of the flood plain of the Semliki River and Lake Albert; and
  • The adjacent Toro plain, a slightly more elevated escarpment area to the east of the flood plain.

Early exploration

The first petroleum geology undertaken in the area was as early as 1920, when Wayland1 described a series of oil and gas seeps at various locations aloprimarily of stacked channel-fill sands, with only very minor amounts of levee and overbank deposits.

Click here to enlarge image

The Kisegi formation has an onlapping relationship to the basement schists and quartzites of the Rwenzori Mountains; therefore, the thickness described at outcrop may grossly underestimate the true thickness of the section. Its most proximal facies adjacent to the basement onlap is conglomeratic, and is known locally as the Kibuku formation, but this facies is not widespread.

Above the Kisegi formation lies a predominantly clay-prone sequence with subordinate sands. These have been subdivided into the Kasande, Kakara, Oluka, Nyaburogo, Nyakabingo, and Nyabusosi formations, based on subtle changes in sand:shale ratio, supported by macrofossil content. However, these formations can be regarded as a single sedimentary unit in which there is some evidence for a cyclic nature, from fluvial sands at the bases of the sequences, through lake margin and ephemeral lake sandstones, silts and clays, to 'permanent' lake conditions dominated by clays.

Click here to enlarge image

The cycles seem to have an average thickness of about 15 m. Only in the basal part of the Nyaburogo formation outcrop are significant sands (30 m) developed at the base of one of these cycles.

1998 seismic survey

In 1998 Heritage acquired an initial reconnaissance seismic survey over the Semliki flats and the adjacent Toro plain.

In order to correctly position lines for this land seismic program, all available field and gravity data were examined, and possible fault patterns were defined by recognition of lineaments observed from satellite data using a Landsat TM image. A good degree of confidence in the derived surface fault pattern is given by the close match between faults identified at outcrop and the lineaments from Landsat.

The derived lineament pattern was used to plan the 1998 seismic grid to give best imaging of faults. Conveniently, it was possible to align the dip lines parallel to the earlier gravity profiles (Fig. 1).

The Heritage seismic survey of 1998 (HOG-98) was the first seismic data acquired in Uganda and lies some 700 km from the nearest African survey on Lake Tanganiyka. A total of 170 km of data was acquired by IMC Geophysics, employing a dynamite source.

The data are of excellent quality to below 4 sec two-way time (Figs. 3 and 4). In particular, down to around 1 sec TWT high frequency reflections of up to 90 Hz provide excellent structural resolution in a complex normal fault setting. Below 1 sec TWT there is excellent reflection continuity throughout the Miocene section. Also a strong basement reflector reveals pre-rift compressional structures (Figs. 3 and 4).

Geological model from seismic interpretation

Interpretation of the seismic data confirms the presence in the basin of a thick section of post-rift sediments that probably represent the Middle Miocene to Recent section described at outcrop (Fig. 2), plus more than 1 sec of additional section of probable Early to Middle Miocene age, that is older that anything described at outcrop.

Below this Late Tertiary rift section is an earlier, possibly Early Tertiary section, separated by an angular unconformity, by analogy to the eastern arm of the East African Rift Valley in Kenya.5 This supposed Lower Tertiary section rests on basement with a variable seismic character, suggesting that both crystalline rocks and old sediments, possibly Karoo, may be present.

The Lower Tertiary and older section is cut by a series of predominantly reversed faults testifying to a pre-rift compressional phase. These faults do not penetrate into the rift section, but act as the sites of propagation for the stems of north-northeast trending flower structures, which are prominent on dip-oriented sections (Fig. 3). The presence of flower structures, that have been active throughout Late Tertiary rifting to the present day, confirms that the area of Block 3 lies at the site of a lateral transfer zone that offsets the rift.

True extensional faults associated with rifting are not interpreted from the seismic data, suggesting that any structure formed by such rifting may be restricted to basin margin areas that were not surveyed.

The major fault within the area, the Makondo fault, forms the eastern flank of the flower structure. It is the same fault described on the earlier gravity profiles, bisecting the basin and forming an escarpment at surface separating the Semliki flats from the more elevated Toro plain.

The Makondo fault runs southwards into the Kibuku fault on the northern flank of the Rwenzori Mountains (Fig. 1). The degree of structuration is high, particularly within the flower structure.

The seismic interpretation suggests that the latest structuration was a transpressionally-induced shortening phase. Evidence for this shortening is more obvious than any extension in the area, and it appears to be associated with the uplift of the Rwenzori Mountains and the adjacent Tertiary outcrop along a series of reversed faults, that has persisted until the present day.

Post-seismic studies

Preliminary basin modeling involved stacking velocities to generate a depth model and using a heatflow model based on field measurements. A heatflow of between 110 and 150 mWm2 had been established from the published data recorded in rift valley transfer zones.6

Heritage has confirmed this geothermal model by analyzing the silica concentration of 'hot' spring water samples collected in the field, using the method of Fournier and Rowe.7 The basin model suggests that much of the section identified from the seismic data is older than the Kisegi formation and is likely to have been in the peak oil or late oil phase of generation during the past 3 MY.

Geochemical analyses of the oil seep at Kibuku (Fig. 1) show that the oil was generated from a nonmarine source rock, with a vitrinite reflectance of 0.65 Ro, dominated by plant kerogen but with some algal affinities.

The GCMS plot (Fig. 5) shows a very high pristane peak, which is typical of nonmarine oils. Based on this information it seems likely that the source rock for the Kibuku oil seep, close to the Kibuku fault, lies beneath the Kisegi formation.

Apatite fission track analysis suggests that the Kisegi formation at outcrop has been buried to more than 1,000 m. Even though previously buried, and subsequently uplifted, the Kisegi formation sandstone retains excellent porosity, ranging between 36% and 39% and permeability ranging between 300 md and 17,000 md.

Based on thin section work, the rock is relatively uncemented and is predominantly quartz with the clay mineral jarosite making up only 5% of the rock. The sand lacks calcite cement and has not been sufficiently buried to develop quartz overgrowths or pressure solution features at grain boundaries.

The Albert graben is very remote from any previous paleontological study areas and no biostratigraphic framework exists for the basin. All previous dating of the section has relied on macrofossils and radiometric dates. The section at outcrop has therefore been sampled to give a reference set from which a biostratigraphic and/or chemostratigraphic framework can be derived to assist, in particular, with stratigraphic control during drilling.

Albert prospectivity

Although the Albert graben represents a new basin play, it also has the exciting potential to be a significant hydrocarbon province, as all the necessary elements of a valid petroleum system are present in the basin.

The Kisegi formation sand represents a world-class reservoir. Field work has shown the sand is in excess of 100 m thick, and analyses have shown that it has extremely good reservoir parameters, which are unlikely to be significantly downgraded at the depths of burial anticipated for currently recognized targets. Also, at outcrop the sands are described as stacked channel fill complexes, which are likely to have good lateral continuity.

The Kisegi formation sand is the primary reservoir target for Block 3.

The sand unit developed at the base of the Nyaburogo formation is thinner (30 m) but also has excellent porosity and permeability, and therefore is likely to provide a good secondary target.

The bulk of the overlying section is highly mud-rock prone, particularly the Kasande formation lying immediately above the Kisegi formation sand. These rocks are likely to provide excellent top and lateral seal to any trap.

Initial interpretations suggest that structural closures exist at Kisegi formation level within the flower structure in the hanging wall of the Makondo fault, and in the corresponding footwall of the fault. It is anticipated that these structures will be confirmed as viable drilling targets when the 2001 infill seismic data are processed and interpreted.

Oil seeps at surface demonstrate that oil is being generated in the basin. Also there is no seismic or field evidence for volcanism in the area, which could be responsible for localized source rock maturity. It seems, therefore, that the source rock responsible for the seep oils is developed regionally.

The Butiaba Waki-1 well penetrated more than 200 m of shale, described as bituminous, which was reported to have total organic carbon (TOC) of 6%, interbedded with sands stained with asphaltic oil.

This shale underlies a substantial sand thought to be equivalent to the Kisegi formation. The age of this shale is unknown, but it has been compared to the Upper Jurassic Stanleyville formation of the Cuvette Central in Congo (former Zaire) but equally may be of Tertiary age.

Whatever the affinities of this shale, it appears to be a valid source rock and may well be responsible for sourcing the oil at the Kibiro seep. More significantly, if this shale is present basinwide, then it may also be responsible for sourcing the oil at the Kibuku seep to the south, suggesting that it constitutes the key source rock horizon for the whole of the Albert graben.

Basin modeling suggests that oils will initially migrate vertically into the Kisegi formation sand that will probably act as a carrier bed, and will then migrate laterally and updip into the prospective areas.

This modeling also confirms that the pre-Kisegi formation section is marginally mature in the kitchen area that is likely to have generated the oil at the Kibuku seep. However, much of the pre-Kisegi section in the north and west of the area covered by the seismic survey is in peak or late oil phase. The extent and thickness of the potential source kitchen mapped from seismic data in the north suggest that volumes of oil expelled are more than capable of filling any of the prospects and leads thus far mapped.

Furthermore, there is evidence of direct hydrocarbon indicators in the seismic data, such as bright (high amplitude) flat spots and possible phase changes, which support the presence of hydrocarbons in the system.

Potential reserves

Interpretation of the HOG-98 seismic data and subsequent mapping has identified a substantial prospect on the footwall of the Makondo fault.

The Semliki prospect (Fig. 6) has mapped closure at both the estimated level of the Kisegi formation sand (2,100-2,250 m), and at the estimated level of the basal Nyaburogo formation sand (circa 1,000 m). Potential reserve volumes have been calculated to be 300-450 million bbl of oil for the primary Kisegi target, and 150-250 million bbl for the secondary Nyaburogo target. The upside potential for a successful dual-target objective for this prospect, therefore, is approaching 1 billion bbl of oil. Early indications suggest considerable potential for this basin.

2001 seismic survey

The 230 km infill seismic program was acquired toward the end of 2001 (HOG-2001), again by IMC Geophysics, employing a dynamite source.

Robertson Research processed the data together with the HOG-98 data that are being reprocessed using the same parameters and LVL information. This new grid will allow the confirmation and improved definition of any previously mapped prospects and leads, and will hopefully allow the identification of further prospects and leads.

The interpretation of these data is currently under way in order to define the optimum prospect that will provide the location for the initial exploration well, in addition to defining further prospects that could provide additional follow-up drilling locations.

Drilling preparations

The Eagle Drill CDU rig has been mobilized to Uganda and is stacked in Kampala. The spudding of the first exploration well to be drilled in the Albert graben in modern times is scheduled for August 2002.

Next week: Meeting exploration challenges in Uganda's remote Albert graben.