1 downloads 42 Views 9MB Size Report
2,096 km of 60-fold Vibroseis seismic reflection data. The location of the seismic data was dictated by two factors— the gravity survey defined limits of the basin ...

Morley, C.K., R.A. Day, R. Lauck, R. Bosher, D.M. Stone, S.T. Wigger, W.A. Wescott, D. Haun, N. Bassett, and W. Bosworth, 1999, Geology and Geophysics of the Anza Graben, in C.K. Morley ed., Geoscience of Rift Systems—Evolution of East Africa: AAPG Studies in Geology No. 44, p. 67–90.

Chapter 4

Geology and Geophysics of the Anza Graben C.K. Morley Department of Petroleum Geoscience University of Brunei Darussalam Negara Brunei Darussalam W. Bosworth Marathon Petroleum Egypt Ltd. Maadi, Egypt

R.A. Day R. Lauck R. Bosher D.M. Stone S.T. Wigger W.A. Wescott D. Haun N. Bassett Amoco Production Company Houston, Texas, U.S.A.

Abstract The Anza Graben is a northwest-southeast trending Cretaceous-Paleogene rift system. The oldest known section from the graben comes from a well in the Chalbi Desert area (northwest area), and is of Neocomian age. It provides a rare glimpse of carbonates in a lacustrine setting. Elsewhere, deposits are dominated by Late Cretaceous-Paleogene lacustrine shales and sandstones, and fluvio-deltaic sandstones. The rift geometry appears to have changed considerably with time—with overall rift activity younger to the southeast. This is clearly seen on the northeast margin boundary fault (Lagh Bogal Fault), which in the central Anza Graben is characterized by predominantly Late Cretaceous activity and in the southeastern graben by primarily Paleogene activity. The key agent affecting basin geometries was the changing activity and location of major faults. A striking feature of the Paleogene tectonic activity, in the southeastern Anza Graben in particular, is the presence of numerous large inversion anticlines which lie sub-parallel to the rift axis. These structures appear to have grown episodically during the Paleogene, alternating with periods of extension.


Chalbi Desert. Despite the latter, shallow sub-surface volcanics considerably degrade the seismic data quality. The three wells drilled by Amoco in the desert yielded important stratigraphic information. In particular, the Sirius-1 well provided the only information obtained to date on Early Cretaceous stratigraphy. The southeastern area is larger and covered by more regionally extensive seismic data. During the 1980s seismic reflection data over the north and central part of the area was acquired by Marathon, Mobil, and Total (formerly the area of exploration block 9), and to the south by Amoco and Petrofina (formerly exploration block 3) (Figure 2). Even older seismic reflection data had been acquired by Chevron in the 1970s. Overall, the data is of much better quality than the Chalbi Desert area (northern graben) because the shallow sub-surface lava flows are generally absent. The area provides important information on how long-lived rift systems evolve.

The Anza Graben is a well developed Cretaceous-Paleogene rift about 500 km (310 mi) long and up to about 130 km (80 mi) wide. Virtually no trace of the rift structure exists at the surface, which is either low-lying arid to semiarid land, or rugged terrain formed by Pliocene-Quaternary lava flows. The graben was first identified from gravity data (e.g., Reeves et al. 1987, Figure 1) and today seismic coverage of the graben is incomplete, hence gravity data remain the only way of defining the entire basin (e.g., Dindi 1994). The Marsabit volcanic province splits the seismic data coverage of the graben into two areas (Figure 2). To the northwest, is a relatively small area encompassing the Chalbi Desert, previously explored for hydrocarbons by Amoco. The surface volcanics presented a serious impediment to seismic data acquisition, hence seismic grid lines were chosen predominantly across the dunes and playa lake of the



Morley et al.

Figure 1. Simplified regional gravity map of the Anza Graben based on exploration data.

CHALBI DESERT Introduction Geological investigations by Amoco geologists included mapping the Chalbi basin using aerial photographs, Landsat, and outcrop data (Figure 3). Access to the area by field vehicles was greatly improved by the presence of the seismic lines. Surface expressions of structure were investigated and stratigraphic sections measured. Fieldwork established that the exposed sequences were fluvial Tertiary sandstones and Pliocene lavas (also see Nyamweru 1986 and Key 1987)—no Cretaceous rocks were identified. The main exposures of sedimentary rocks contain fluvial sandstones of the Maikona Formation. Thin, interbedded red and green shales are poor in microfossils, but micropaleontological results suggest a Miocene to Pliocene age for the formation.

Paleocurrent indicators showed sediment transport directions were generally axial to the Anza Graben (i.e., from northwest to southeast). Surface lineaments identified from aerial photographs suggested that faults were oriented parallel to the basin axis and predate the Maikona Formation. However, clastic dikes and associated joints indicate minor north northeast-south southwest extension persisted in the Chalbi Desert region during deposition of the Maikona Formation (Marathon unpublished data, Bosworth and Morley 1994—their Figure 3). Even today, the north northwestsouth southeast oriented dune fields and playa lake loosely parallel the underlying structural grain (Figures 3 and 4). The surface geological information was integrated with 2,096 km of 60-fold Vibroseis seismic reflection data. The location of the seismic data was dictated by two factors— the gravity survey defined limits of the basin and the areas

Geology and Geophysics of the Anza Graben

of extensive lava flows. Not only did the thick lava flows at the surface significantly reduce data quality, they also made acquisition very slow and expensive. The lava flows cover the deepest parts of the basin. The two main horizons that could be mapped were firstly, an event thought to be near the top of the Lower Cretaceous and secondly, the top of Precambrian basement (see Figures 4 and 5). The Lower Cretaceous reflection is the most continuous event in the data. However, in areas of poor data quality the continuity of the reflection is highly questionable. Several anomalous events exist within the data set and appear to cross-cut the reflections from sedimentary rock packages. Some are due to igneous dikes and sills (confirmed by the wells Sirius-1, see Figure 6, and Chalbi-3), others may come from diagenetic boundaries within the sedimentary units.

Structure of the Chalbi Desert Gravity data provides the only complete picture of the geometry of the northwestern Anza Graben (Figure 1). It shows that the Anza Graben narrows considerably towards the northwest and dies out on the eastern side of Lake Turkana. The northwest-southeast trend of the northern margin of the Anza Graben is very consistent. The rift narrows considerably along a north-south trend on its south-


western margin. The deepest part of the basin, as indicated by the largest negative gravity anomaly, lies below the surface volcanics and probably corresponds to a series of major (southwestern dipping) boundary faults that form the northeastern margin of the rift. The Chalbi Desert area forms a relatively high area flanked to the west, north, and east by deeper basins (Figures 4 and A1). Seismic data confirmed the presence of a complex, structurally high area in the subsurface of the Chalbi Desert, later tested by the Sirius-1 and Bellatrix-1 wells. These data also imaged the half graben basin lying to the west of the high. Figure A1 illustrates one of the better quality seismic lines from the area.

Stratigraphy of the Chalbi Desert Sirius-1 Well Sirius-1 was Amoco’s first wildcat well in the area (Block 10). It reached a total depth of 8,656 ft (2,838 m) (Figure 6). A 3,000 ft (1,000 m) thick section of Tertiary sandstones, below near-surface volcanics, was encountered before reaching Cretaceous sandstones and shales. The sedimentary section was cross-cut by two Late Cretaceous intrusions which form prominent seismic reflectors in the area, as predicted by the synthetic seismogram (Figures 5 and 6b). Two major unconformities were identified, one

Figure 2. Distribution of volcanics, wells, and seismic lines in the Anza Graben (mostly after Bosworth and Morley 1994).


Morley et al.

Figure 3. Geological map of the Chalbi Desert compiled from Landsat, aerial photos, and fieldwork. See Figure 2 for location.

where Turonian clastics overlie Neocomian clastics and limestones, the other where Miocene sandstones overlie early Campanian sandstones and shales. The Sirius-1 well is important because it contains the only Lower Cretaceous rocks identified in the Anza Graben. The presence of Lower Cretaceous age rocks demonstrates the geological similarity between the Anza Graben and the rift basins in Sudan (Schull 1988, Bosworth 1992).

Early Cretaceous - A total of 2,120 ft (695 m) of Neocomian age section is interpreted to be present in the well. Lithology, electric log ,and geochemical data suggest the following environments: shallow water lacustrine; lake-margin eolian; and transgressive lacustrine. The shallow water lacustrine environment is marked by a 187 ft (61 m) carbonate mudstone. It overlies a late Miocene igneous intrusive at total depth. The carbonate

Geology and Geophysics of the Anza Graben

section is characterized by: 1) a consistently high gamma ray response ranging from 80–120 API units; 2) a capping 11 ft (3.5 m) thick organic shale zone (3.6% TOC); and 3) a variety of carbonate textures consistent with lake margin environments. The thin shale was dated as Neocomian by the presence of the pollen Dichieropollis. No marine indicators were observed. Carbonate textures observed from cuttings included organic muds, ooids, pellets, algal thrombo-


lites, concretions, iron-stained reworked clasts, and coquinas containing ostracod and pelecypod fragments. A wavy texture resembling microstromatolites, but of questionable origin, was also observed. The textures suggest a shallow water environment enriched in calcium carbonate and experiencing fluctuations in lake level. The concretions and iron-stained clasts indicate a paleosol formed by subaerial exposure on a lake margin. The presence of ooids and

Figure 4. Time structure map of the Chalbi Desert area at approximately top Early Cretaceous. See Figure 2 for location.


Morley et al.

Figure 5. Detailed interpretation of seismic line TVK-21 in the vicinity of the Amoco Sirius-1 and Bellatrix-1 wells (see Figure 4 for location of line).

coquina suggest a swash zone environment, while the organic-rich zone may represent a lacustrine marsh deposit. These sub-environments are interpreted to have existed in a lake margin position similar to that of the modern carbonate environments observed in Lake Tanganyika, offshore Burundi. There, seven carbonate sub-environments described by Cohen and Thouin (1987) parallel 80 miles of shoreline and are distributed according to water depth. Overlying the carbonate with a sharp contact is a 478 ft (126 m) sandstone. It is a massive, moderately sorted, medium- to fine-grained quartz arenite. The gamma ray log is anomalously “clean” compared with the sandy section shallower in the well (Figure 6). Few, to no, shale beds exist. Electric log porosites range from 15–22% over 40% of the section and permeabilities vary from 10–100+ md. These characteristics suggest an eolian sand. Stratigraphic analysis of the dipmeter log supports the possibility of an eolian environment, however, this interpretation is not exclusive. Ten thin dune-forms ranging from 5–20 ft (1.5–6 m) thick can be identified. They display decreasing dips from 50°–0° with depth. These zones are surrounded above and below by generally chaotic dips. An alternative environment is that of a lacustrine shoreface. The sharp contact between the carbonate and the sandstone is interpreted as a lacustrine regressive event that resulted in subaerial exposure of the lake margin carbonate and marsh environment. Dune sands were then deposited. The situation is analogous to the southwestern shores of Lake Turkana today where dune sands extend for over 70 mi (112 km) overlying older carbonate lake deposits. Quaternary alluvial sands also overlies algal carbonates in the Kargi area of the Chalbi Desert (Figure 3). Shales (streaks) interbedded with sandstone form the boundary between the eolian sandstone and the overlying 1,455 ft (443 m) lacustrine transgressive unit. The shales

are probably interdune deposits formed during a rise in lake level that reworked the dunes. Above 7,855 ft (2,395 m) the shales become more abundant and sandstones more lithic and less quartz-rich. Lacustrine conditions gradually became dominant, depositing less mature sediments (compared with the reworked dune sands). The remainder of the Lower Cretaceous is represented by lacustrine environments with interbeded sandstone and shale ranging in thickness from 5–40 ft (2–13 m). Late Cretaceous - Palynological data indicate that the Aptian-Albian section is absent in the well, hence an unconformity between the Early and Late Cretaceous exists. Palynomorphs have provided only general age ranges for the Late Cretaceous section, with a possible range from Turonian to early Campanian. A fluvial-lacustrine environment for the Turonian to early Campanian rocks has been identified. The section consists of 3,020 ft (920 m) of interbedded sandstones and shales, similar in character to the lacustrine transgressive unit in the Lower Cretaceous. The presence of Pediastrum bifidites var “longispinum” and rare ostracod fragments suggest a freshwater lacustrine environment. Tertiary - Below 618 ft (188 m) of Pliocene shallow volcanics is 2,652 ft (808 m) of Miocene to Recent alluvial to fluvial sediment. This age is based on the presence of palynomorphs “no older” than Miocene. The section is clean on the gamma ray log and shale-poor compared with the Cretaceous section. Source rock - Five thin source-quality shales (TOCs > 2%) were identified in the Cretaceous section. They range from 3.4–9% TOC over intervals 10–33 ft (3–10 m) thick and contained amorphous kerogens. However, even the richest zones had low convertibility, casting doubt on their ability to produce significant quantities of hydrocarbons. The presence of minor amounts of oil in the well recovered during RFT testing indicates the presence of a mature oil source

Geology and Geophysics of the Anza Graben

rock somewhere in the basin. Following an analogy with the rift basins of Sudan (Schull 1988), the source rock could be Aptian-Albian age shales lying on the flanks of the high tested by the Sirius-1 well. Vitrinite reflectance determinations on 20 sidewall cores and picked cuttings in the richest zones show that the sec-


tion above 3,380 ft (1,108 m, Miocene to Recent) is thermally immature and the section below (to total depth, Cretaceous) is in peak maturity for hydrocarbon generation. Spore coloration indices (SCI) are in general agreement with the vitrinite data. Anomalously high vitrinite readings occurred in three zones above igneous rocks in the well. These baked zones provide the main evidence for their intrusive, rather than extrusive, origin. Bellatrix-1 Well The stratigraphy of the Bellatrix-1 well consists of extrusive volcanics, Pliocene, Miocene, and early Campanian to Cenomanian rocks (Figure 7). The entire section is clastic and appears to be non-marine due to the absence of any marine fauna and the presence of freshwater algae. The Miocene-early Campanian unconformity represents a gap of about 56 million years. The Miocene section is 6,937 ft (2,114 m) thick, nearly twice that of the Miocene in the Sirius-1 well. It is interpreted to be an alluvial-fluvial sequence based on the thin to absent shale beds (clean gamma ray curve) and coarse sands. The lower part of the section expands into an extensional fault that was active at the time of deposition (Figure 5). The age of the section indicates extension broadly contemporaneous with the Kenya Rift. Late Cretaceous rocks with no identifiable depositional breaks are present from 6,937–11,414 ft (2,114–3,479 m, total depth). They are characterized by 10–100 ft (3–30 m) thick sandstones interbedded with shales of comparable

Figure 6. a. (left) Lithological column and well logs for Amoco Sirius-1 well (gray = predom. Ss., horizontal dashes = predom. Sh.) b. (above, right) synthetic seismogram for the Sirius-1 well. Note that the igneous units form the most prominent reflections.


Morley et al.

Figure 7. Lithological column and well logs for Amoco Bellatrix-1 well (gray = predominantly Ss., horizontal dashes = predominantly Sh.).

Figure 8. Well logs for Amoco Chalbi-3 well.

Geology and Geophysics of the Anza Graben

thickness (Figure 7). The section from 6,937–9,900 ft (2,114–3,050 m) probably ranges from early Campanian to Turonian in age and correlates to the 3,380–6,400 ft (1,030–1,951 m) section in the Sirius-1 well. From 9,900–11,414 ft (3,050–3,479 m) the sedimentary rocks are of Cenomanian age. There is a suggestion from both wells that the time period, of approximately 10 million years, between the Cenomanian and Campanian is poorly represented and may reflect a depositional hiatus or possibly uplift and erosion. More data is required to resolve the cause. In both the Sirius-1 and Bellatrix-1 wells the Miocene and Upper Cretaceous sandstones are texturally immature arkoses and litharenites, strongly suggesting they were derived from a nearby source and experienced similar transport histories. Scattered Albian-Aptian palynomorphs present between 7,200 ft (2,195 m) and total depth were interpreted as having been reworked. They provide evidence that Albian-Aptian sediments existed or exist somewhere in the Anza Graben (Bosworth 1992, Bosworth and Morley 1994). Between 7,439–7,469 ft (2,267–2,277 m) a conventional core was acquired. The objectives were to obtain stratigraphic and source rock information black shale and coal beds, and to core across the interval that corresponded to the “green” seismic reflector (near the Cretaceous-Tertiary boundary). The cored section contains interbedded dark gray, coaly shales and fine-grained sandstones, overlying a massive sandstone at the base. The section is interpreted to represent deltaic and shoreline environments on the margin of a rift lake. Vitrinite reflectance (Ro) values from sidewall cores ranged from 0.6 at 7,270 ft (2,216 m) to 1.07 at 11,130 ft (3,448 m). Values measured from cuttings ranged from 0.47 %Ro at 7,350 ft (2,240 m) to 0.66 %Ro at 11,250 ft (3,429 m). The values obtained from cuttings were probably lower due to caving and sample mixing. Overall, the vitrinite reflectance data shows the Cretaceous section to be at early peak oil generation. Pyrolysis data are consistent with the Ro and SCI data, showing Tmax values increasing with depth to 160°C (320°F) at total depth. Chalbi-1 Well The Chalbi-1 well reached a total depth of 11,953 ft (3,644 m). It penetrated Tertiary-Cretaceous sandstones interbedded with thin intervals of fluvio-lacustrine siltstones, claystones and shales (Figure 8). Four igneous intrusions were crossed. Porous sandstone was encountered from the surface to the igneous intrusion at a depth of 5,190 ft (1,582 m). Below this sandstone porosity degraded severely with depth due to cementation by zeolites (particularly laumontite). The well is located in a more basinal position than either the Sirius-1 or Bellatrix-1 wells. It penetrated a relatively thin Tertiary section (2,010 ft, 612 m) and a thick Upper Cretaceous section (6,290 ft, 1,917 m). The Maastrichtian was not identified in the well and the Cretaceous section begins in the late Campanian. From 8,300–11953 ft (2,530–3,644 m) the well encountered the elusive Aptian section that is not present in any other well in the basin—


although reworked Aptian-Albian microfossils are common in other wells. Overall, the well increases in shale content with depth.

Subsidence Modeling The goal of subsidence modeling is to match an observed vitrinite reflectance profile (in a well) with a calculated profile from a model based on rock/sediment ages and a reasonable burial history. For the Bellatrix-1 well, the observed thermally mature Cretaceous section occupies a 4,000 ft (1,219 m) interval. The bottom of the well does not enter the gas generation window. A relatively long (2,500 ft, 762 m) oil window was simulated by the modeling (Figure 9). However, it was not possible to match the observed thickness. An alternative solution was to increase the thickness by decreasing the geothermal gradient below the present day measured gradient, however, this was not considered reasonable. The calculated vitrinite reflectance slope closely matches the observed slope for Bellatrix-1 well. The modeling suggests that the structurally high complex tested by the Sirius-1 and Bellatrix-1 wells experienced a significant amount of burial and subsequent erosion, probably during the early Paleogene. It appears that the missing section was at least 8,000 ft (2,439 m) thick in the vicinity of the two wells. The earlier erosional event or events, which removed the Aptian-Albian section, were of insufficient magnitude compared with the effects of later burial to produce a “jump” in maturity between the Late and Early Cretaceous sections (Figures 9b and 9c).

SOUTHEASTERN ANZA GRABEN Basin Stratigraphy The first well drilled in the vicinity of the southeastern Anza Graben was the Wal-Merer-1. It was drilled on a gravity high and tested a Tertiary and marine Cretaceous section. Subsequent exploration in the mid-seventies was based upon gravity and seismically defined leads and resulted in the exploration of the Tertiary section by the Anza-1, Bahati-1, and Meri-1 wells. None of these wells were located in the Anza Graben, although the Anza-1 well lies close to the southeastern termination of the Anza Graben gravity trend (Figures 1 and 10). Fluvio-lacustrine rocks were predicted to occur in the graben, this was confirmed when Total, Mobil, and Marathon drilled the N’dovu-1 well in 1987 and encountered a predominantly Late Cretaceous continental section which had experienced episodic marine influxes (Winn et al. 1993). Four wells have been drilled in the central and southeastern Anza Graben: Hothori-1; N’dovu-1; Kaisut-1; and Duma-1 (Winn et al. 1993, see Figure 10 for well locations). All four reached total depth in Late Cretaceous age rocks. Hothori-1 Well The Hothori-1 penetrated the thickest section of the four wells with a total depth of 4,390 m (14,400 ft, Figure 11). The lowermost 1,200 m (3,900 ft) are of Campanian age and


Morley et al.



c. Figure 9. Thermal maturity modeling for the Amoco Bellatrix-1 and Sirius-1 wells. a. (above, left) Bellatrix-1 well, match between vitrinite reflectance data (Ro) maturity. There is a jump in maturity between the present day geothermal gradient and that calculated for the Upper Cretaceous (from vitrinite reflectance data). This requires deposition, uplift and erosion of about 8,000 ft (2,400 m) of section sometime between the late Cretaceous and late Tertiary. b. (above, right) Sirius-1 well, the high vitrinite reflectance values, like those in the Bellatrix-1 well also require deposition, uplift, and erosion of about 8,000 ft (2,400 m) of section sometime between the Late Cretaceous and late Tertiary. c. (right) Burial history plot for the Sirius-1 well illustrating two periods of uplift and erosion. They probably represent the end of rifting episodes (one in the early Late Cretaceous, the other either during the latest Cretaceous or early Tertiary). The magnitude of deposition, uplift, and erosion for the Late Cretaceous-early Tertiary is derived from Figure 9b.

consist of immature, often red-colored sandstone, mudstone and shale, and dark gray lacustrine shales with thin sandstone beds. This sequence is overlain by several sandstone and siltstone packages separated by minor disconformities and unconformities, all probably of early Tertiary age. The uppermost 600 m (1,970 ft) of the well correspond to the unfaulted sequence shown in Figure 12 of Miocene to Recent age which has a slight marine influence. N’dovu-1 Well The lower part of the N'dovu-1 well is dominated by medium gray to black shales that contain rare dinoflagellates and reworked palynomorphs of Albian-Aptian age (Figure 11). These rocks are of Cenomanian age and appear to represent deep water lacustrine and brackish water deposits that episodically underwent marine incursions. The Campanian/Maastrichtian section in the well is composed of siltstone and poorly sorted, immature sandstones. The Tertiary section is thin compared with the Hothori-1 well and overlies the Cretaceous section with a marked angular unconformity (Figure 13).

Both the N'dovu-1 and Hothori-1 wells were drilled in the main trough of the Anza Graben. The graben’s subsidence was controlled by the northeast side boundary faults (including the Lagh Bogal Fault). The Kaisut-1 and Duma-1 wells were drilled in half graben trends that lie to the west of the main depositional trough (Figure 10). Duma-1 Well Duma-1 was drilled north of the Matasade Horst block, on the western flank of the main Anza Trough (Figures 11 and 14). The well reached a total depth of 3333 m (10,933 ft) in probable late Cretaceous shales, siltstones, and poorly sorted sandstones. The Cretaceous is capped by approximately 300 m (980 ft) of Tertiary sandstone and conglomerate, probably of Miocene and younger age. The early Tertiary appears to be absent at this well. Kaisut-1 Well The Kaisut-1 well reached a total depth of 1,450 m (4,760 ft) in early Tertiary or latest Cretaceous age poorly cemented sandstone with minor shales (Figure 11). Seismic

Geology and Geophysics of the Anza Graben

data show that a section of steeply rotated, high-velocity rocks lies below well total depth. They may correlate with Upper Cretaceous strata at Sirius-1, or may be older Cretaceous or even Jurassic strata (Bosworth and Morley 1994).

Structural Geometry The largest negative gravity anomalies associated with the Anza Graben show a northwest-southeast trending trough, with the steepest gradients occurring on the northeastern margin (Figure 1). They mark the location of the Lagh Bogal Fault and other boundary faults. On the western side of the rift two gravity high areas stand out and correspond to horst blocks. Smaller half graben basins on the west side of the rift can also be identified from the gravity data. The gravity contours associated with the rift do not open out into the Lamu Embayment to the south, but instead close northwest of the Anza-1 well. This discontinuity is significant in a geological sense because the Anza Graben, despite its proximity to marine section to the southeast, is dominantly composed of continental rocks with only episodic marine incursions. There is considerable structural complexity to the Anza Graben with marked along-strike changes in geometry. On regional dip cross-sections the graben can be sub-divided into the northeastern deep Anza Trough, and the southwestern shallower region of tilted fault blocks (Figures 1, 10, 15, A2, and A3). It is convenient to describe the alongstrike variations in these sub-regions separately. Anza Trough This area contains at least 9 km of Cretaceous to Recent sedimentary rocks. The northeastern margin of the Anza


Graben is bounded by major boundary faults along its entire length. However, in the vicinity of seismic line 86-140 to just south of line 86-160, the boundary faults have considerably reduced displacements, and for much of the history of the basin’s edge this area was a weakly faulted flexural margin (Figure 16). In the southwestern area of the trough the expansion of section into the faults is dominantly Paleogene (Figure 17, A4, and A5). There is little evidence for significant expansion of Cretaceous-aged section into the faults, except during the Campanian (see Chapter 8, Figures 16 and 17). Consequently, the southeastern portion of the Lagh Bogal Fault is a dominantly Paleogene feature and separate half graben bounding faults operated during the Cretaceous (Figure 17). Passing southeastwards, Tertiary depocenters become younger (Figure A4). One marked feature of the seismic data across the Anza Trough is the presence of a central, poor data quality zone. The N’dovu-1 well was drilled on the flank of this zone in the footwall of a tilted fault block. The extensional fault bounding the structure appears to have been largely active during the Cretaceous and was inverted during the Paleogene (Figure 16, see Chapter 12). Probably much of the problem with poor quality seismic data along this zone is due to the inversion. On seismic line 86-160 the inversion structures are relatively subdued, they become more pronounced to the southeast (Figure 18). On the southwestern margin of the basin one or more low-angle faults appear to have been active during the Cretaceous and Paleogene (Figure 16). One very large listric fault gave rise to a large rollover structure (whether it is detaches within the sedimentary section or passes into basement is unsure) and may have accommodated some 20

Figure 10. Generalized map of the central and southeastern Anza Graben showing the location of key seismic lines and wells.


Morley et al.

Figure 11. Lithological columns for three wells in the central and southeastern Anza Graben (see Figure 10 for well locations), gray = sandstone, horizontal dashes = shale.

km of extension. This low-angled geometry is a characteristic of the northeastern margin of the intermediate tilted fault blocks (see below) and the Matasade High (Figure 18). To the southeast, the basement-involved character of the normal faults changes as does the structural style (Figures 15 and 19). The overall deep geometry of the rift remains, but the main Paleogene depocenter changes from the southwestern to the northeastern margin as the low-angle faults on the western side die out and the Lagh Bogal Fault zone increases in displacement (Figure 19). Dip cross-sections across the Anza Trough show a series of faulted anticlines in the Paleogene-Upper Cretaceous section (Figures 15, 18, and A5–A7). They typically display

eroded crests broken up by planar and listric normal growth faults, which detach at one of several horizons within the Cretaceous section. In the Hothori-1 well one of these detachment horizons is a slightly overpressured lacustrine shale of early Campanian age (Figure 12). The deep fault pattern is semi-independent of the anticline-growth fault pattern (Figure A5). The anticlines are commonly asymmetric and do not display the correct fault-fold geometric relationships to be considered simple rollover anticlines. The best interpretation seems to be an initial growth fault structure developed by gravitational instability within the sedimentary section during rifting that has subsequently been subject to alternating phases of extension and inversion

Geology and Geophysics of the Anza Graben

(see Chapter 12 for a more detailed discussion). The strong detachment between the upper and lower sedimentary sections (apparently due to Upper Cretaceous lacustrine shales) resulted in decoupling of some inversion structures from basement-involved inversion faults. A time structure map of the Anza Trough is shown in Figure 19. The mapped horizon is approximately the top Campanian. It illustrates the general northwest-southeast structural grain that affects both faults and folds. The anticlinal closures are mostly inversion features. The absence of the stepped, en echelon fault and fold geometries, and the orientation of many normal faults sub-parallel to the fold axes are not features typically associated with strike-slip anticlines. However, the alternation of episodes of extension and inversion, plus the effects of inherited fabrics make it difficult to discern whether compressional or strike slip motions created the inversion structures. Intermediate Tilted Fault Block Province This province lies between the Matasade High and the Anza Trough. It is comprised of a series of closely spaced tilted fault blocks within sedimentary rocks of predominantly Cretaceous age (Figures 14, 15, and 16). The Duma1 well has shown that the deformed section is of Maastrichtian-Campanian age at the top—the older section is of unknown age, but is probably Late Cretaceous. There is no indication of strong Paleogene extension affecting the province, however inversion features (probably of Paleogene age) are present in some areas. The Duma-1 well has higher seismic velocities for a specific depth than do other wells in the Anza Graben. They range from 3,000–5,000 m/sec (9,800–16,400 ft/sec). Sandstones in the well were highly compacted and cemented and the penetration rate of the drill bit was very slow. The well has very high vitrinite reflectance values and the data suggest paleo-temperatures in excess of 120°C (250°F) have affected the entire section. Uplift and erosion proba-


bly occurred during the Paleogene, when the western boundary fault to the Anza Trough was active, and the intermediate tilted fault block province (in its footwall) underwent isostatic uplift and possibly inversion. Later igneous activity may have expelled fluids responsible for the precipitation of laumontite cements. Kaisut Basin The western half of the Anza Graben is characterized by numerous low-angle normal faults, relatively thin basin fill, and several distinct angular unconformities (Figure 20). The dominant faults have changed significantly with time. The northern part of the area (seismic line 86-175, Figure 20f) is characterized by two deep, simple half grabens that thicken westwards into east-dipping boundary faults. Towards the south the basin fill thins, boundary faults become dominantly west- to southwesterly-dipping, and several angular unconformities are present. In a particular occurrence, a highly rotated lower section (Cretaceous) is separated from a less rotated upper section (Paleogene). However, the exact timing of the unconformity is unsure. Local unconformities associated with inversion during the Paleogene are widespread, as well as a later Tertiary unconformity (Miocene?) that marks the end of most tectonic activity (see Figure 7 in Chapter 8). In the southern part of the area the cut-off angle between the west-dipping faults and the reflections from sedimentary units is large, indicating they were originally high-angle faults that have been rotated by movement along other faults to a lower angle (Figure 20). These faults are cut through by a later, higher angle east-dipping fault (the Kaisut Boundary Fault). The unconformities described above mark significant periods of basin reorganization, as the activity on faults changed. This can be seen in the map view evolution of the basin in Figure 20. During the Cretaceous the Kaisut Basin was composed of a southeastern province dominated by

Figure 12. Seismic line 74-10, with correlation to the Hothori-1 well (see Figure 10 for location).


Morley et al.

Figure 13. Correlation between the N’dovu-1 well and seismic line 86-160 (see Figure 10 for location). The well is situated on an uplifted footwall block, that was eroded at the end of the Cretaceous to early Tertiary.

west to southwest-dipping normal faults, that passed northwestwards into a province dominated by northeasterly-dipping faults (Figure 20, also see Figure 7 in Chapter 8). Paleogene deformation smoothed out these differences to some extent by the southeastward propagation of the northeasterly-dipping Kaisut Boundary Fault. In contrast, the westto southwesterly-dipping faults became inactive or had much reduced displacements. Inversion along the northeasterly-dipping faults produced some well developed anticlines late in the basin’s evolution. The only well in this area is the Kaisut-1 well, unfortunately all the samples analyzed were barren of microfossils

Figure 14. Correlation between the Duma-1 well and seismic line 86-145A (see Figure 10 for location). Most of the faults have a simple normal fault geometry, however, the anticline at the northeastern end of the line suggests some late inversion occurred. The strongest reflections come from shaly units. The generally poor, discontinuous reflections indicate most of the section is sand-rich.

and no age dating was possible. The section is dominantly sandstones with shale intervals. It may be entirely Tertiary, or the lower part of the section may be Cretaceous. During drilling the rate of penetration was fast, indicating a poorly cemented and undercompacted section.

Comparison of Interval Velocity Profiles in the Central and Southeastern Anza Graben Interval velocities for basinal areas in the central Anza Trough, southeastern Anza Trough, and Kaisut Basin are compared in Figure 21. The different regions show significantly different interval velocity profiles that reflect their

Geology and Geophysics of the Anza Graben


Figure 15. Geological cross-sections made from depth-converted seismic lines. Correlation of geological ages away from well control is difficult and therefore speculative. Note how the uppermost Late Cretaceous section in the Anza Trough becomes buried under thicker Paleogene section passing southeastwards. Seismic evidence for (deeper) section below the uppermost Late Cretaceous rocks becomes weaker passing southeastwards. This may be a problem of seismic resolution or it may reflect younger rift initiation to the south with time. On seismic lines 86-125 and 74-10 inversion features are prominent (see Figure 10 for locations).

differing geological histories. The central Anza Trough displays a gradual increase in interval velocities with depth that is relatively unremarkable except for comparative purposes with the other areas. In the southeastern Anza Trough, within the same basin, the same simple velocity profile does not exist (Figure 21b). Instead, the profile can be broken into two, with a marked offset to the deeper velocity profile, at about 2–2.5 seconds two way transit time (TWTT). This depth corresponds approximately to the location of the base of the Paleogene and top of the Campanian sections. This marked increase in velocity across the unconformity indicates a change in rock properties possibly the result of deeper burial followed by uplift and erosion of the Cretaceous section, or a Cretaceous age diagenetic event. When the central and southeastern Anza Trough data are plotted together the section below two seconds (TWTT) plots in the same location for both areas. The shallower thick Paleogene section in the southeastern Anza Graben has significantly slower interval velocities than the Cretaceous section at corresponding depths in the Central Anza

Graben. The Kaisut Basin displays consistently higher interval velocities for a particular depth than the Anza Trough (Figures 21c and d). This suggests that in general the basin fill is more highly cemented and has less porosity than the Anza Trough. Hence, the Duma-1 well may be more representative of the Kaisut Basin than is the Kaisut-1 well. The origin of the porosity reduction is unclear, but the data suggests that the history of fluid migration in the Kaisut Basin and adjacent Anza Trough was very different.

Rift Flanks On the eastern margin of the Lagh Bogal Fault Amoco drilled two stratigraphic tests in the footwall section (Figure 1). The wells reached total depths of 1,281 m (4,200 ft) in the Elgal-1 well and 2,051 m (6,730 ft) in the Elgal-2 well in Karroo age sedimentary rocks. The stratigraphy was similar to sections in outcrop further to the northeast—primarily quartzitic alluvial fan and braid plain deposits. An important shale unit approximately 80 m (260 ft) thick occurs at the top of the Permian. In the Elgal-1 well the


Morley et al.

Figure 16. Geological evolution of seismic line 86-140 (see Figure 10 for location). The line illustrates that the Lagh Bogal Fault was not a major fault until the Paleogene. It also shows that the western bounding fault to the Anza Trough (at about the 65 km mark) developed into a major fault during the Paleogene.

Figure 17. Map illustrating changing fault activity with time in the central and southeastern Anza Graben.

Geology and Geophysics of the Anza Graben


Figure 18. Detailed interpretation of western side of Matasade High area, seismic line 86-130, illustrating inversion features (see Figure 10 for line location).

Figure 19. Time structure map on the top Campanian in seconds, Anza Trough.


Morley et al.


Figure 20. Tectonic evolution of the Kaisut Basin as seen in map view and on seismic data (a.–f.). Series of dip seismic lines (migrated) across the Kaisut Basin demonstrating the marked northwest-southeast changes in structural geometry (see time structure map, Figure 19, and Figure 10 for line locations). The change from northeastern reflection dips into southwest-dipping faults, in the southeastern area, to southwestern stratal dips into northeastern-dipping faults in the northwestern area is very marked. In the southern area some older faults terminated at the Cretaceous-Paleogene boundary. This, coupled with the development of an important northeast-dipping boundary fault on the southwestern margin of the basin during the Paleogene, reflects an important reorganization of basin geometry during the Tertiary. It shows that the initial rift geometry of alternating fault orientations can be smoothed with time to produce a larger region with more homogeneous basin and fault orientations.


Geology and Geophysics of the Anza Graben






Morley et al.


Figure 20. (cont.) Refer to figure caption on page 84.

Figure 21. Comparison of interval velocities vs. depth (two way travel time in seconds), for different regions of the Anza Graben: a. Summary of the data for all the areas; b. southeastern Anza Trough; c. central Anza Trough; and d. Kaisut Basin. Gamma ray curves for the wells are after Winn et al. (1993).

Geology and Geophysics of the Anza Graben

present day geothermal gradient was calculated as 1.78°F/100 ft (3.24°C/100 m), while for Elgal-2 well it was 1.37°F/100 ft (2.49°C/100 m). Cuttings and sidewall cores were examined for spore coloration and vitrinite reflectance. A sample from the Elgal-1 well analyzed for spore color yielded a TAI value of about 6. Vitrinite reflectance values were not reliable, but were estimated to be about 2. In the Elgal-2 well a sidewall core at 790 m (2,592 ft), a combined core sample from 1,701 and 1,708 m (5,581 and 5,604 ft) and a single sample from 1,725 m (5,659 ft) contained sufficient organic material to yield vitrinite reflectance values of 1.72, 2.00, and 2.02%Ro respectively. The thermal maturation data and high sonic velocities, 4,900–5,550 m/sec (16,100–18,200 ft/sec), in the Karroo section indicate that it has undergone considerable burial followed by uplift. However, only a broad estimate of the uplift amount can be made—in the range of 1,600–3,000 m (5,200–9,800 ft). Reflection geometries of the Karroo section from the seismic line the wells were drilled on (line 742), indicate that the depth to basement at the northwestern end of the line is about 8,400 m (27,600 ft), at the Elgal-1 well it is 5,600 m (18,400 ft). The difference between the two depths suggests that approximately 2,800 m (9,200 ft) of uplift has occurred in the section penetrated by the wells. Approaching the Lagh Bogal Fault, the base Karroo reflection rises to within a few hundred meters of the surface indicating uplift of the order of 8,000 m (26,200 ft) has occurred over a distance of about 60 km (37 mi). This magnitude and wavelength of uplift is much greater than would be expected for flexural-isostatic footwall uplift and suggests an important thermal or magmatic underplating event sometime between the end of Karroo deposition and early Miocene. The timing of uplift could have coincided with the mid-Jurassic rifting of Madagascar from East Africa, or occurred during the Cretaceous-Paleogene rifting. The presence of Karroo rocks along the footwall of the Lagh Bogal Fault suggests that a considerable thickness of Karroo section might underlie parts of the Anza Graben. There is, however, no indication of this section from wells or seismic data. The Cretaceous rift basin section is too thick and masks any older basins that might be present. The same can also be said for the Jurassic section. Along the Matasade High are some thin deposits of quartzites and carbonates. The carbonates are probably of Jurassic age. It is very likely that during the events that led up to the rifting away of Madagascar from Kenya, a Jurassic triple junction existed in the area of the Lamu Embayment and the failed arm extended northwest along what is now the Anza Graben (Reeves et al. 1987). The carbonates are the only evidence for this scenario—again drilling and seismic data have failed to identify the presence of a Jurassic rift. It is therefore impossible to assess the impact that older structures might have had on the Cretaceous rift system at this time.

SUMMARY The large size, structural complexity, and sparse well and seismic data available limit current understanding of the basin’s evolution. A regional paleoenvironmental view of


the evolution is presented in Figure 22, but it is highly interpretive. A general scheme for the timing of tectonic and sedimentary events is given in Figure 23. The authors suspect that to fully understand the evolution of the Anza Graben would require deep drilling—beyond depths of current economic interest. The main events that have affected the Anza Graben are as follows: Neocomian - The Neocomian section has only been penetrated by Sirius-1 well. The sequence contains carbonates of probable lacustrine origin overlain by eolian sandstones, deposition is thought to have been associated with a lake margin. It is possible that the Kaisut Basin (along strike to the southeast) was also active during the Neocomian. Age dates obtained from the Duma-1 well are controversial— either Early or Late Cretaceous (Winn et al. 1993). Following Winn et al. (1993), a Late Cretaceous age is the preferred interpretation. Aptian-Albian - This section is absent from all wells in the Anza Graben except the Chalbi-3 well. A marine section of this age is known from the Wal-Merer-1 well in the northern Lamu Embayment. The Aptian-Albian was an important time for lacustrine source rock deposition in the Sudanese Cretaceous rifts (Schull 1988) hence it was considered important for exploration to establish the presence of like rocks in the Anza Graben. The Sirius-1 well contains Upper Cretaceous rocks that unconformably overlie Neocomian rocks and several wells contain reworked Aptian-Albian palynomorphs. Aptian-Albian age rocks are inferred to exist in the deeper parts of the basin. Oil shows in the Sirius-1 well could have come from a lacustrine (Aptian-Albian) source rock down-dip. The age of the erosional event is not well established and could range from within the AptianAlbian sequence to early Cenomanian. The erosion may have marked the end of an early phase of rifting. Cenomanian - The central Anza Trough contains several hundred meters of Cenomanian (?) deep water, marine, brackish and lacustrine shales in the N’dovu-1 well. This suggests at least episodic opening to the Indian Ocean via the Lamu Embayment. In the northwest, the Cenomanian rift-fill sequence is characterized by coarse-clastic fluvial deposits. Input of these clastics is believed to have occurred dominantly by transport from the northwest along the axis of the basin. Correlation of the seismic data between the Hothori-1 and N’dovu-1 wells indicates that almost all the basin fill of the main Anza Trough can be attributed to sediments of Late Cretaceous and younger age. A deeper, older Cretaceous section might be present but there is little evidence for it. In the basin evolution diagrams of Figure 22, the Anza Graben is shown initiating rifting during the Neocomian only in the north. Between the Cenomanian and Campanian there are only poorly dated, thin sections that possibly represent the intervening time period. This suggest there was an important break in extensional activity sometime between the Turonian and Santonian stages. Campanian - Coarse fluvial sandstones, overbank deposits and episodic lacustrine shales occurred throughout most of the Anza Graben during the Campanian. In the cen-

Figure 22. Evolution of the Anza Graben based on well and seismic data (redrawn from Bosworth and Morley 1994).

88 Morley et al.

Geology and Geophysics of the Anza Graben


Figure 23. Timing of tectonic and stratigraphic events in the Anza Graben.

tral Anza Trough eroded Early Cretaceous sediments were redeposited. In the southeast, the Hothori-1 well penetrated more than 300 m (1,000 ft) of early Campanian lacustrine rocks. The Campanian age section is the one time period that is found in all parts of the rift system and probably marks an important phase of extension. Maastrichtian - Microfossils and traces of anhydrite in the Duma-1 well suggest a brief marine incursion during the Maastrichtian. Coarse fluvial clastics continued to dominate sedimentation in the rift system and fault activity was reduced. However, the Maastrichtian section appears to have been better developed in the vicinity of the N’dovu-1 well than the Campanian section (Figures 11 and 13). The important topographic features of active rifts appear to have been smoothed and a regional axial drainage system developed. In the southeastern Anza Graben there is no evidence for the presence of any Maastrichtian section (Figure A4). The absence of any marked unconformity suggests that tectonic activity in the southeastern area was minimal and basins were not formed at this time. Tertiary - The structural evolution of the Anza Graben during the Tertiary was complex and displays considerable along-strike variation in timing and structural style. Early Tertiary deposits are unknown from the area northwest of Mt. Marsabit. Although they are thin in the N’dovu-1 well, the Paleogene section appears to thicken to several kilometers on seismic data into a northeast-dipping boundary fault on the southwestern margin of the Anza Trough. In the southeastern part of the trough the section thickens to over 3,000 m (10,000 ft) adjacent to the Lagh Bogal Fault. Normal faults were also active in the Kaisut Basin during the Paleogene - In the Chalbi Desert area there was probably an episode of major erosion during the Paleogene, at least in

the vicinity of the Sirius complex, where approximately 2,440 m (8,000 ft) of rock has been eroded. During the Paleogene multiple inversion events affected the central and southern parts of the Anza Graben, intermittently interrupting extension. In the central and southern Anza Graben Miocene and younger deposits are commonly less than 300 m (1,000 ft) thick and form a flat-lying unconformity (commonly angular) indicating a post-Paleogene cessation of significant tectonic activity. The Miocene section in the Hothori-1 well contained microfossils showing significant marine influence. Since the Miocene the basin has been thermally subsiding. In the northwest area Miocene age units are up to 600 m (1,800 ft) and appear to be associated with late fault activity—this is best documented in the Sirius-1 and Bellatrix-1 wells.

REFERENCES CITED Bosworth, W., 1992, Mesozoic and early Tertiary rift tectonic in east Africa: Tectonophysics, v. 209, p. 115–137. Bosworth, W., and C.K. Morley, 1994, Structural and stratigraphic evolution of the Anza rift, Kenya: Tectonophysics, v. 236, p. 93–115. Cohen, A.S., and C. Thouin, 1987, Nearshore carbonate deposits in Lake Tanganyika: Geology, v. 15, p. 414–418. Dindi, E.W., 1994, Crustal structure of the Anza Graben from gravity and magnetic investigations: Tectonophysics, v. 236, p. 359–371. Key, R.M., 1987, Geology of the Marsabit area: Kenya Mining Geology Department report, No. 108, 42pp. Nyamweru, C.K., 1986, Quaternary environments of the Chalbi basin, Kenya: Sedimentary and geomorphological


Morley et al.

evidence, in Sedimentation in the African Rifts , L.E. Frostick, R.W. Renault, I. Reid and J.-J. Tiercelin eds., Geol. Soc. London. Special Publication, v. 25, p. 297–310. Reeves, C.V., F.M. Karanja, and I.N. Macleod, 1987, Geophysical evidence for a failed Jurassic rift and triple junction in Kenya: Earth and Planetary Science Letters, v. 81, p. 299–311.

Schull, T. J., 1988, Rift basins of interior Sudan: petroleum exploration and discovery: AAPG Bulletin, 72, p. 1128–1142. Winn, R.D., J.C. Steinmetz, and W.L. Kerekgyarto, 1993, Stratigraphy and rifting history of the Anza Rift: AAPG Bulletin, 77, p.1989–2005.