<<

Downloaded from http://sp.lyellcollection.org/ by guest on September 30, 2021

African and sedimentation, an introduction

H.G. Reading

Rifts are one of the most spectacular features of Gulf of Elat rift lies along a that global morphology. Within oceans they separate separates two lithospheric plates and, as Arabia plates as new oceanic is created. Within moves northwards relative to the Palestinian they may form deep valleys such as the (Mediterranean) block and faults curve or side- Rhine , within which runs one of the step each other, extensional pull-apart basins are world's busiest waterways. They are often floored formed, such as the whose surface lies by deep , such as Baikal, the deepest further below than any other place in the lake in the world today whose floor lies 1700 m world. A feature of these strike-slip rift valleys is below the surface of the lake. Ancient are the that the extensional valleys pass, along the rift, sites of petroleum accumulations beneath the into mountainous zones of compression and margins of the Atlantic, in the and in uplift. Rapid erosion from these leads China, of coals and oil shales and of , to substantial sediment supply for the basins from including phosphates, barite, Cu-Pb-Zn sul- within the system itself. Ancient strike-slip phides and uranium (Robbins 1983). Of all the rift basins are therefore characterized by very world's rifts none can match in scale and diversity thick, rapidly accumulated basin fills with areas the Great Rift System, which runs for over of contemporaneous compressional tectonics and 7000 km and includes the unconformities not far away. System and its extension through the Rift basins that are extensional features into the Dead Sea, to form a unique feature of on a regional scale tend to sink rather more global geology. Its study is important not only for slowly, have less sediment entering them, and what it can tell about the nature and origin of show no evidence of synchronous compressional present-day rifts, the thermal, magmatic and tectonics in their neighbourhood. Cross-sections tectonic processes which gave rise to them, clima- through them show normal faulting, probably tic changes, and sedimentary, particularly lacus- listric in form. trine and volcaniclastic, processes. It is also The main controversy with regard to extension- essential for the understanding of the processes al rift valleys is whether magmatic activity in rifts which formed passive continental margins, all of reflects an underlying heat source such as a which are underlain by rifted sedimentary basins, convection plume in the or hot spot, the and failed rifts whether they formed within rifting then being the consequence of expansion continents or at the junction of continents and and uplift brought about by thermal activity, or oceans. whether the magmatic activity is merely the result Early workers on rifts were impressed by the of regional or local extensional stresses which wide valleys, 40-50 km across, filled by young have passively allowed to come to the alluvial and lacustrine sediments and bordered by surface. steep escarpments rising in some cases a few Some rifts that penetrate continents are clearly hundred metres, in others up to 3-4000 m above failed rifts or that began as one arm of the valley floor. Many rifts are the sites of a . As ocean floor spreading deve- volcanic activity, sometimes of rather unusual loped along two arms, one arm was left as a 'failed composition. These unusual volcanic source rift' in the sense that new was not rocks may in turn be the cause of peculiar lake created. Perhaps the best known is the Benuc water chemistries, many rift lakes containing trough within which the developed, unique suites of evaporitic minerals. However, the northern (Viking) graben of the Although at one time some people considered North Sea has been claimed to be one (Whiteman rifts to have been formed by compression (ramp- et al. 1975), and so has that part of the East valley hypothesis), discussion today revolves African Rift known as the where thc principally around the extent to which rift valleys Red Sea and the Rifts successfully are extensional features or are the result of drifted, whilst the Afar triangle 'failed'. oblique-slip movements between two laterally Molnar & Tapponnier (1975) have argued that moving blocks. other rifts are the result of , in There is no doubt now that some rift valleys are particular that the impact of India with the result of strike-slip movement. The Dead Sea/ caused not only major lateral displacement along

From FROSTICK, L.E. et al. (eds) 1986, Sedimentation in the African Rifts, Geological Society Special Publication No. 25, pp. 3-7. Downloaded from http://sp.lyellcollection.org/ by guest on September 30, 2021

4 H.G. Reading faults such as the Red River and Altyn Tagh different from the West and Central African Rift strike-slip faults, but also E-W extension to System, and its southern and older portion is very produce the Baikal graben. The Rhine graben has different again from the northern Kenyan to also been interpreted as the result of stresses Ethiopian parts. The dominant feature of the generated in the Alps. East African Rift System is the wide zone of low However the rifts in are not the result of density which is expressed topo- continental collision, and there is little evidence graphically by the localized Ethiopian and for any substantial degree of strike-slip move- Kenyan domes where volcanic activity is most ment. However, strike-slip movement may occur active. In the northern portion the rifting appears locally and some African rifts lie along major to be relatively young (late Tertiary) though there zones or extending for thou- is evidence that rifting commenced during the sands of kilometres, which may have had a latest (Williamson et al., this volume). considerable amount of lateral movement at The amount of crustal extension is relatively some time in the past. small, about a quarter of that in the West and Although the East African Rift System is the Central African Rift System (I0 35 kin, Fair- best known of African rifts, other extensive rifts head, this volume), and there seems little doubt occur especially in West and that the rift is the result of thermal activity in the (Fairhead, this volume). Some parts have been upper mantle and consequent thermal uplift. well studied, for example the of Support for thermally-induced rifting comes because of its petroleum resources, and from the migration of some rifts eastward with the Volcanic Line is also well known. preceding doming, that in turn pre- Other parts, though, have little topographic cedes rifting. The reason it has not opened yet expression and therefore have only recently may simply be that regional stresses are not yet become known, due to hydrocarbon exploration. conducive to opening. Were they to be so it would In particular the nature of the geology across the probably open. Central African has not been under- Sedimentation in the northern portion is com- stood. monly away from the domes and main rift. For This complex pattern of shear zones, asso- example the is to the W of the ciated rifts and rift valleys, known as the West present active rift, though it probably started and Central African Rift System (Fairhead, this within a late - rift before the rift volume), is apparently older than the East Afri- migrated eastward (Cerling & Powers 1977). can Rift System in that at the present day it is Sedimentation may also take place between rifts largely inactive except for some gentle subsi- in sag basins such as . Sedimentary dence. The main rift phase was the thicknesses are not too great; volcanic accumu- when substantial and rapid stretching of the lations are considerable, and volcanic source caused thinning of the crust and areas give rise to unusual and varied sedimentary passive upwelling of hot . It deve- compositions. loped as a result of large extensional stresses, The southern portion of the East African Rift continental disruption and high sea-floor spread- differs from the northern portion in that volcanic ing rates. Volcanism was minor and although activity is limited. Although there is still a wide, there was probably a large anomalously low low density zone on either side of the rifts, there is density body within the upper mantle during the no thermal doming. The initiation of rifting Cretaceous it did not produce domal uplift, appears to be earlier, and rifts follow Precam- except perhaps at the oceanward end of the Benue brian lineaments. Sediments in the rift basins are Rift. Instead deep rifts formed, open to the sea. thicker and up to 3-4 km of sediments may have This was due to the major crustal extension which accumulated. occurred at this time, perhaps some four times From the early days of the that of the East African Rift System. Towards the hypothesis, the East African Rift has been consi- end of the Cretaceous some lateral displacements dered to be a present-day example of the earliest took place along the Central African Shear Zone, stage of continental breakup, that is the rift stage and during the Tertiary wider and more extensive which precedes the drift stage represented by the sedimentary basins than the original rift basins Gulf of Aden and the Red Sea (e.g. Drake 1972). formed as the original thermal anomaly decayed It was then supposed that in a relatively short and the lithosphere contracted. Unfortunately, geological time the East African Rift would itself apart from the Benue trough, little is known open and a new ocean would be created along its about sedimentation in these -related rift length. basins. This view, however, was opposed by those who The East African Rift System is significantly pointed to the relatively long-lived nature of some Downloaded from http://sp.lyellcollection.org/ by guest on September 30, 2021

African Rift tectonics and sedimentation parts of the East African Rift, stressing in asymmetrical basins c. 50-70 km long and 20- particular structural evidence in the exposed 40 km wide. This alternation of two overlapping rocks, particularly in the central and and interfacing sub-basins can result in appar- southern portions of the rift (McConnell 1980). It ently symmetrical basins with a in the was argued that the whole 5000 km length of the middle being seen in some cross-sections. In plan African Rift System, from the Afar triangle to the view has a sinusoidal pattern Orange river in South Africa, was a Precambrian resulting from this overlapping of arcuate or taphrogenic lineament which has since been fre- crescent-shaped sub-basins. The sedimentologi- quently reactivated. On this view the East African cal effect of the division of the lake into sub-basins Rift is a more or less permanent feature within a separated by is that for the most part the which is unlikely to be split apart. lake is divided into a number of individual To some extent both views are correct since depocentres more or less independent of each there is evidence in some part of Africa that: (1) other, probably with reversed facies transitions rifting frequently takes place along lines of older except where spillover can occur between sub- lineaments which have long geological histories; basins. (2) the type of movement and rifting reflects Ancient volcaniclastic sediments are notor- global stress patterns; (3) mantle processes can iously difficult to interpret. The normal models of cause both linear and radial zones of high heat sedimentation used to interpret sediments formed flow, low density upper mantle, volcanic activity in non-volcanic terraines no longer seem to apply. and topographic uplift, as in the Red Sea, Gulf of The reasons are: (1) the diverse, compositionally Aden and the northern East African Rift; and (4) immature, volcanic source rocks; (2) the rapid global stresses on their own, without heat sources -tectonic movements that cause complex from the upper mantle, may themselves form patterns of uplift and and rapidly rifts, as for most of the West and Central African changing water levels; and (3) the build-up of Rift System. volcanic edifices, both subaerial and subaqueous, As with all linear or curvilinear features on the on the scale of large mountains thousands of globe, rifts are not the continuous simple features metres high to small, but sedimentologically they appear to be when looked at from a satellite. important, flows. These lava flows can, in They are invariably broken into segments by one instant of time, transform the topography transverse features. The larger the scale at which and form both physiographic barriers which they are seen, the more complex they become. isolate depositional systems and temporary local Individual segments vary considerably in their sources of sediment. structural style. This segmentation is probably a As shown by Cohen et al. (this volume), a prime factor in controlling sedimentation since it spectrum of rift models can be postulated, based controls the rate of basin subsidence and the on the degree of volcanism and rift subsidence. relative subsidence within the basins, and the Where the rate of volcanism exceeds that of rift amount of sediment supply and its composition. propagation rifts are totally filled by volcanics Although most rift models were initially deve- and their derived sediments. Where there are loped as essentially symmetrical models, deepen- moderate to high extrusion rates relative to rate ing towards the centre, it has gradually become of rift propagation, shallow lakes will form. apparent in the past few years from detailed Where volcanism is absent or minor, deep rift studies and geophysical profiles that continental lakes develop. Rifts can thus be divided into: (1) rifting is an asymmetrical process (see Bally those that are volcanically filled, in which lakes do 1983). This is the consequence of rift geometry. not exist, e.g. the , N of latitude 1~N However, not only are rifts asymmetric in cross- (Williams & Chapman, this volume); (2) volcani- section, but the direction of asymmetry appears cally-dominated rift basins with lakes, e.g. the to vary systematically along a rift with the Turkana ; and (3) non-volcanic rift direction of asymmetry alternating between seg- basins such as Lakes and Tanganyika. ments. Between each segment are transverse Non-volcanic lakes are almost devoid of lake structures called transfer faults, transform zones margin sediments, most sediment being trans- or accommodation zones. In some cases these ported direct to the lake centre. However, in zones are ridges and they may inherit pre-existing volcanically-dominated lakes volcanic barriers structures. form both across the main rift dividing it into sub- Many areas of the East African Rift System are basins, and parallel to the lake margin to form known to display segmentation from detailed sediment traps where lake margin sediments mapping (e.g. Williams & Chapman, this accumulate, particularly coarser-grained mater- volume). In Lake Tanganyika seismic profiles ial as around . In non-volcanic (Rosendahl et al., this volume) show alternating basins Cohen et al. (this volume) predict there Downloaded from http://sp.lyellcollection.org/ by guest on September 30, 2021

6 H.G. Reading should be a higher proportion of coarse clastic volcanism, climate is almost as important. It not deep water turbidites. only governs the amount, rate and periodicity of Thus the study of volcanism and its effects in runoff, and the composition of the sediment rifts helps towards the understanding of ancient supplied, but also influences the environment of volcaniclastic sediments, the rapidly changing deposition in the lakes and interfluvial areas that sedimentary environments and the influence of make up so large a part of the depositional areas varied source areas. In addition, another effect of of rifts. In these environments, the productivity volcanics, as pointed out by Eugster (this and nature of the plants and animals living in volume), is that a volcanic floor to a basin is often water or forming soils, and the chemistry of extensively fractured and very porous. Conse- inflowing surface and spring waters and of lake quently streams are short and much water is waters, are crucial to sedimentation. For this supplied to the lakes by perennial springs, rather rcason also, ancient rift sediments are one of the than from surface flow. best indicators of palaeoclimate. Work on East A final consequence of volcanism is ash fail. African rift sediments by Pleistocene strati- Firstly, ash falls are the most reliable and short- gvahers has given us a measure of Pleistocene lived chronostratigraphic markers that exist in climatic changes matched only by research oil geology. Therefore, a fc,~turc of wHcanically- glacial deposits and, more dominated rifts is the abundance of good strati- recently, on oceanic sediments. Fortunately lhe graphic markers which nlake complex sedimen- Vast African Rift System crosses all the tropical tary facies patterns intelligible. Secondly, subacr- environments, I'rom the high rainfall equatorial ial ash falls can, as with lava flows, fi~rm I~uild-tlps bell, through the fluctuating climates of Ihc of considcrable thickness. lhirdlv. and probably stl bl topical zones, to the undoul~tcd deserts 1o the the most imporlandy, they preserve . N and the S. A range of sediments can thus be including footprints and sofl bodies, in places sludied, flom those forming in freshwater, biolo- where fossils do not normally form. In continen- gically very productive lakcs, to those forming tal environments fossils are m~rmally only pre- under conditions of diverse salinity and composi- served at lake margins, in anoxic lakes whcrc l~on in both peremfial and cphcmcral lakes. plants and animals can sctllc undisturbed and bc For lhc sedimentologist, the virtues of studying preserved from oxidation; as detrital particles lakes are manitbld, especially where the sources (bones and teeth) in fluvial sediments; and as of the sediments are so varied. Lakes are micro- burrows or traces in palaeosols or peat build-ups. cosms of larger marine basins. Within a relatively Volcanic ash, however, preserves faunas in all sma I1 area one can study rivers, deltas, shorelines, types of environments (Bishop 1968) including and shallow and deep water sedimentation. The . Of particular importance is the proximity of active normal faulting allows one to chemistry of the ash (Hay, this volume; Pickford, compare sediments forming at the present day this volume). The optimum conditions are those with those uplifted a few thousand or more years of low silica and high lime. East African volca- ago, within a tectonic and sedimentary context noes are famous for their , and most not so different from that which is currently of the eruptive magma was probably operating. alkali rich or natrocarbonatite. As pointed out by The Lake Turkana basin, for example, is now Hay (this volume), when natrocarbonatite ash is one of the best studied lake basins in the world. altered by contact with water it can provide: (1) The lake is the largest (7500 km 2) closed-basin an instant cement of sodium carbonate (= trona); lake and the fourth largest natural lake in Africa. (2) an alkaline environment in which bone apatite In the past it has had a much larger extent. Several and calcium carbonate fossils are relatively inso- papers in this volume describe the processes luble; (3) a fine-grained matrix for preserving involved in the interaction of sediment supply footprints and moulds of soft-bodied organisms. from short-lived ephemeral streams, a major including the muscles of birds and even the delta, and shoreline and cliff erosion, with the tongue of a lizard; and (4) a supply of calcium slightly saline (2500ppm) and alkaline lake carbonate for calcification and coating of bones, waters (plt 9.2) no deeper than 125 m. replacement of plant material and filling of lhanks largely to the work of Eugster and his bones. It is no coincidence that is the colleagucs (Eugster, this volume) and its use as a domain of so many beaulifully preserved conti- comparison with lhe Green River Formation in E~er~tal animals (particularly primales and other Ihc I lnilcd States, Lake Magadi is, to geologists, mammals, and reptiles), especially soft-bodied the best known African lake. It is a moderately ones. salinc, alkaline ephemeral closed-basin lake Although sedimentation in rifts is primarily which has served as a model for the formation of controlled by tectonics and the degree and type of sodium carbonates and Magadi-type cherts in Downloaded from http://sp.lyellcollection.org/ by guest on September 30, 2021

African Rft tectonics and sedimentation 7 many ancient sedimentary rocks. Of increasing lakes, stromatolites are known from many basins importance, however, is (Renault in the African Rift System, even if some types are et al., this volume; Vincens et al., this volume). not forming today (Casanova, this volume). They Though similar in many ways to Lake Magadi, form as lacustrine stromatolites, as oncolites and Lake Bogoria is perennial and is fed not only by as travertine chimneys and pool-rim dams, each hot springs but by seasonal streams. It is saline with its own characteristic morphology. Onco- (>40,000 ppm) and highly alkaline (pH 10.3), lites grow on floodplains during inundation, and and within its basin a range of carbonates form. travertines of several distinct forms develop at Non-marine carbonates, though at one time springs, at waterfall dams and in rivers. sadly neglected, are now established as a major There is thus no of the world to match lithology that occurs extensively in ancient sedi- African lakes for the study of the biochemical mentary rocks. Although lacustrine carbonates processes in lakes, especially alkaline lakes within are not widely developed in alkaline soda-rich a range of volcanic and tectonic backgrounds.

References BALLY, A.B. 1983. Structural styles and the evolution of MCCONNELL, R.B. 1980. A resurgent taphrogenic linea- sedimentary basins. Am. Assoc. Petrol. Geol., Educ. ment of Precambrian origin in eastern Africa. J. Short Course 198 pp. Geol. Soc. 137, 483-9. BISHOP, W.W. 1968. The evolution of fossil environ- MOLNAR, P. & TAPPONNIER, P. 1975. Cenozoic tectonics ments in East Africa. Trans. Leicester Lit. Philos. of Asia: effects of a continental collision. Science Soc. 62, 2244. 189, 419 26. CERLING, T.E. & POWERS, D.W. 1977. Paleorifting ROBBINS, E.I. 1983. Accumulation of fossil fuels and between the Gregory and Ethiopian Rifts. Geology metallic minerals in active and ancient rift systems. 5, 4414. 94, 633 58. DRAKE, C.L. 1972. Future considerations concerning WHITEMAN, A., NAYLOR, D., PEGRUM, R. & REES, G. geodynamics. Bull. Am. Assoc. Petrol. Geol. 56, 260- 1975. North Sea troughs and plate tectonics. Tectono- 8. physics 26, 39-54.

H.G. READING,Dept of Earth Sciences, The University, Parks Road, Oxford OXI 3PR, UK.