Jou",~1 0/ A/riet'" £~"h Sri~lI,t$. Vol. S. Nos. 213U. pp. 16.'1 ... 182. 1911* Printed in G r<:3t Britain

11S'I'.l-5)621l1~ SH ill + 0.00 © t%V Pe rgamon Press pic

Is structure the main control of river drainage and sedimentation in rifts?

LYNNE FROSTICK- and IAN REID+

• Royal Holloway & Bedford New College. Egham TW20 OEX. UK + Birkbeck College, London WClE 7HX. UK

Abstract- Changes in the topography of riflS as iheyevolve toward aulacogens and passivc marginsconlrOl river drainage ald. therefore. sedimentation. The African continent offers an opportunity to assess the effect of rifting on river drainage and sedimentation at several stages of development. The East African system is young and sediment-starved as aresult of both the regional doming that accompanies eady phases of rifting and the back-tilting of footwall fault blocks. Each of these factors diverts riVets away from subsiding basins. The longitudinally segmented nature of an early rift discourages axial drainage and major river inputs originate on the hanging-wall platfonn or roll-over of half graben sub-basins. During later stages of development. regional subsidence may enlarge the drainage basin. while sedimentary infilling of the trough may facilitate axial drainage. A good example of this occurs in the West African system where the Benue River nows for 1200km along a rift trough and the drainage basin incorporates the catchment of the River Niger_ Sedimentary sequences accumulating in of! basins are characterised by strong cyclicity of depositional style. This can be taken to reOect the sporadic nallJre of the pulses associated with faulting . Howevet. itcan also arise from the changes in base-level that are brought about by fluctuations in climate in closed basins and by eustatic adjusonenlS in sea-level where marine inundation has occurred. Taking examples from the Eas t Afric.an system and from the Gulf of Suez. changes in sedimentation that are consequent upon marine or lacustrine transgression UK!

regression are shown to be almost indistinguishable rTom those triggered by boundary fault activity. Cyclical patterns of sedimentation can only be attributed to tectonic activity if corroborative evidence is available .

INTRODUCTION

Over the past 5 years. ideas concerning the pattern of sedimentation that occurs In rifts have changed considerably_ Following tl1e work ofBrtdge and Leeder {19791 . models of basin fill have re­flected recent advances tn the general under­standing of rtft structure (e.g. Frostick and Reid 1987a and 1987b; Leeder and Gawthorpe 19871. These advances have occurred because of the Significant expanSion of geophySical exploration that fonowed the oll crises of the late 1960's and 1970's.

Of particular Jrnportance has been the work car­ried out In theTert1aryrifts of East Africa (Rosendahl et aL 1986; Khan el al 1986; Rosendahl 1987). However. African rifts are not confined to the east of the continent. nor to the Tertiary period. There have been sporadic bouts of crustal extension since Pre-Cambrian times (Grove 1986). Moreover. while younger rtfts have occasionally re-used old lines ofwealmess (e_g. the Rukwa rift in Tanzarua). they have often established new fauli trends more In keepingwtth prevaUlng stress fields. as in much ofthe Gregory Rift In East Africa. As a result. Africa offers for study a unique range ofrill development. from the early stages of continental rifting to the evolution of passive margins.

The pattern ofrtverdrainage associated with rifts can be expected to change as structural evolution takes place and as erosion alters the landfonn. Because rivers are major suppliers of sediment. it might also be expected that sympathetic changes would occur in sedimentation. Structural history should exert a strong control on the character of the basin fill. This paperexamlnes the sedimentary consequences of rift development by dra .... 1ng out the similarities and contrasts in drainage patterns and sedlments of three major systems. The Tertiary-Quaternary system of East Africa is young and stut active. In contrast. the older Cretaceous rill of West Africa lS tectOnically quiescent. though sUIl having topographical expression. Both are presently continental In character. On the other hand. the Red Sea has a long history of marine Influence. and Is evolving Into a passive margin .

GEOLOGICAL BACKGROUND

The geological hlStory of the African Rift System is well documented (e.g. Baker 1970: Baker et aL 1972; Falrhead 1986) and only a brief review Is Included here.

The development of the East and West African and the Red Sea RIft systems spans two major periods of continental fragmentation. The first period was associated with the opening of the

165

166 LYNNE FROSTICK and IA N REID

South Atlantic during the late Cretaceous. 1Wo arms of a trilete rtft system evolved to Conn the Bight of Benin. The remaining ann • the Benue Trough - failed to develop in the same way. As a result. tectonic activity died out and the flanking fault scarps have been softened by erosion.

The East African Rift System contrasts with that oeWest Africa in having a long history of volcanism (Baker 1986; Williams and Chapman 1986). This gives, among otherthrngs. a good chronostraugra­phlcal framework for the tnterpretaUon of basin fills (Baker eta!. 1971; Fitch eta!. 1978: Chapman and Brook 1978), Fairhead (1986) suggests that the contrast reflects diiferences in the degree and rate of extension rather than any fundamental differences in mantle structure. Estimates of the age of Inception of the East Afrtcan Rift System range from late Cenomanian to early Miocene (Arambourg 1933; Baker et al 1972: King 1978: Williamson and Savage 1986). It Is associated with a phase of continent fragmentation which was res· ponsible for development of the Red Sed/Gulf of Aden proto-ocean. The Red Sea/Gulf of Aden and East African Rifts meet in the Afar "triangle" of EthIopia. Like Its West Afrtcan counterpart. the East African Rift represents the failed arm of a trilete system, although at a much earlier stage of development.

The Red Sea/Gulf of Aden proto-ocean Is opening at an estimated rate of 10 mm.a- 1 (Courtlllot et aL 1987) as a consequence of the movement of the Arabian plate northeastwards away from Africa. The Gulf of Suez Is a northern extension of this system. with a history of faulting dating back to the Oligocene ( Sellwood and Netherwood 1984). The earliest phases of movement were probably as­sociated more with crustal relaxation following the AJpine orogeny than With inCipient continent break­up . However. later developments of Pllo­Pleistocene age reflect the relative motions of the Arabian and Afrtcan plates. Recent small scale activity IS confined to offshore locations close to the axis of the Gulf(Ganunkel and Bartov 1977). Most of the differential plate movement IS currently being accommodated by strike-slip movement on the Gulf of Aqaba/Dead Sea fault (CourtUiot et at 1987). In consequence. the Gulf of Suez IS also a rift that has failed to develop fully. However. it has a history of marine Inundation which dates back almost to Us inception.

THE INFLUENCE OF RIFTING ON REGIONAL PATTERNS OF DRAINAGE

A major geomorphological and sedimentological consequence of conUnental rifting is the diversion and realignment of river drainage, whatever the pre-existing pattern. The character of the re-

ordered drainage will control, In tum. the pattern and nature of clastic sediment supply to the deve­loping rift basins. It therefore becomes a chief de­tenninant of the architecture of the resulting basin fill. Where major river systems are captured by the rift. thick accumulations of alluvial sediments may develop. In contrast. areas of the rift receiving no fluvtalinputs may become zones In which organlc­rich. carbonate and diatomaceous deposits occur (Cohen and 1110uin 1987: Cohen in press).

One widely held mJsconception is that rills are sumps that always attract drainage from a wide area. The development of facies models In which extensive allUvial fans emanate from the rift foot­walls reflects this idea (e.g. Hubert et at 1976). However. the topographical expression of rift struc­ture. especially duting the early phases of develop­ment, may not encourage dIversion of rivers into the rift. ThIs is the case for all rifts located over mantle plumes. where regional doming is one of the first expressions of inCipient continental break­up. As a result. sediment supply to young rift basins can be highly localised and small-scale (Frostick and Reid 1987 a.b: Leeder and Gawthorpe 1987). In this type ofrift it Is only as the system ages that there Is a tendency for the in­tegration of drainage over a wide area. At this stage clasttc sediment Inpllts can become more evenly distributed about the baSin. but by this time It may be topographically indistinct. as in the case of the Abu Gabra system of southern Sudan.

The trends In drainage development that occur as rifts evolve are evident In the river systems of Afrtca. For this reason, the regional patterns of drainage contributing to the West and East Afrtcan systems and to the Red Sea have been extracted from pubUshed maps. The areas Involved are loea· ted In FIg. I: the drainage nets and major faults are shown in Figs. 2. 3 and 4. In addition, a larger scale analYSiS of the rivers and structure of the 145 000 kml Suez drainage Is shown in Fig. 5.

ImmedJately obvious are the very limited catch­ment areas of the East Afrtcan. the Red Sea/Gulf of Aden and the Suez rtfis (Figs. 2. 4 and 5]. All can be conSidered as young systems. Rivers draining to each of these rtfts have headwaters less than 350 km from the basin axes a10ng their entire length· over 900CHan In total. Much of the local drainage flows away from. rather than towards. the basinS. This is especially evident In the Gulf of Suez (Fig. 5). Here. much of the drainage close to its western side flows towards the Nile, while very little of the Sinai contrtbutes on Its eastern side. most rtvers flowing northeastwards towards the Wadi el Arish • an ephemeral river that drains to the Mediterranean.

So great Is the trend away from rift basins dUring early rifting that Lake Vlctorta. caught between 2

EffeclS of rifling on river drainage 167

F1g. 1. Index map shoWing the areas In which the relatlonshlp betwee n patterns of river drainage and structure are analysed .

A - East African Rift (F1g. 2); B - West African Rift (F1g. 3): C - Red Sea/Gulf of Aden (F1g. 4): o -Gulf of S uez (Ftg 5): E - Lake 1\nkana Basin (Fig. 6).

branches of the East African system (Fig. 2) re­ceives the waters of 8 major rivers, all backshed from the rift shoulders. Indeed. because the topo­graph ic gradient is onen away from the rift (Frostick and Reid 1987a). several of the present lift lakes, notably Tanganyika and Malawi, are starved of clastic sediment. As a result, sedimen­tation rates are low (0.5 mm a·L compared with almost 7 mm a'L forL. Turkana) and both biogenic and carbonate facies are particularly well developed (Cohen and Thouin 1987).

The situation in West Africa Is very dUTerent (Fig. 3). Here the rift is a major drainage corridor reaching towards the Atlantic Ocean. The Benue River runs for most of its course along the basin axis and is Joined by the Immense Niger system which runs ttslast480km within the Benue trough. A thickness of > 5 kin of predominanUy clastic sediment has accumulated since the Cretaceous (OjO and AJakalye 1976) . Pleistocene rates of sedi­mentation have been particularly high. prodUCing

substantial growth of the Niger Delta (Allen 1965. 1970). This Is thought to have resulted from the southward impingement of Saharan dunes. and the transport of thiS easily eroded matertal from the Inland Delta region of the Niger and down the Benuewhlch was, at that time, swollen byoverspill from Megachad (Crove 19861.

STRUCTURAL CONTROL OF DRAINAGE DIVERSION

Early Rifting Phase Diversion of river drainage is probably achieved

very early in the history of rift basin development. In fact. the early phase orrtfting in East Afnca and other. similar , areas situaled over mantle plumes. involves regional doming (McKenzIe 1978: Rosendahl 1987) so that the river courses on pene­planed cratons such as that of Africa are likely to be threatened before substantial rift faulting oc­curs . As doming Increases, even entrenched rivers

168 LYNNE FRosn el( and IAN REID

D D

"E +s s

Or~ inile not issociited WI\h .11\ buin

f,ulu . wHh throw wh. re known

Fig. 2. River drainage and s tructure of the East African Rill. M - MaJagorasi River: 2 and 3 - the locaUons of the Koobi Fora Formation and Kapthurtn Formation sequences. respecUvely. given in Fig. 7.

are going to be affected. Once the rift basins develop. the topographic gradient will exclude rivers from entertng the rtft over a conSiderable portion of its length ILeFoumler et al. 1985: Frostlck and Reid 1987a.bl.

It Is increasingly aclmowledged that the funda-

mental structural unJl of rtfts Is a half graben controlled by movement on a Single. major. l1stric boundary fault (Gibbs 1984 Rosendahl etal 1986). Footwall collapse and the backtilting offault blocks, retnfortes the tendency already establiShed by doming to shed rivers away from the r1.fl basin

Effects of rifti ng on river drainage 169

5 ,

+15 N

I BIGHT OF

BENIN

./ Rivers, perennial and ephemeral

lak es, pe renn ia l and ephemeral

o Drainage to r i ft basin

C·--' Drainage not associated -----1 with rift ba sin

..A' Faults , with throw /"' where known

~ Str ike-s lip faults

--

F1g. 3. River drainage and structure of the West Afr1can Rift. CASZ- the Central Afr1can Shear Zone.

(Frostick and Reid 1987a). Few riveTS and little sediment enters the rift on its faulted margins (see Table 1) . However. on the hanging-wall a gradient is estabUshed towards the baSin. and rtvers which might have previously drained across an uninter­rupted surface are captured by the developing rift basin . For example. the invertebrate fauna of the Malagorasi River (Fig. 2. letter M). which now drains to the eastern shore of Lake Tanganytka.in­dlcates that it was at one time a trtbutary of the

Congo River (A. Cohen. personal communJcaUon). The hanging-wall platfonn or roll-over therefore becomes a major route for rift Jivers. sometimes responsible for over 70% of the total basin drainage area (Table 1) .

Of great consequence for the pattern of sediment supply to rift basins Is the dlscontlnuous nature of the boundary faults . Half graben segments are rarely more than 150 km long and the major boundary fault shuts from one side of the rtft axis

170

MEDITERRANE" SEA

•••••••• . ;; }

LYNNE FROSTICK and I AN REID

D

I!lvt:'>.pc'cnn'.I1 ~nd "phcme,~1

L"kh, perenmal and epheme,,,1

Drainage no. Hso(.ated wllh ,ilt iJa,in

F,uill ), wilh Ihrow whe re known

Suike-slip fauhs

il; Ull

••••

FIg. 4. RIver drainage and structure of the Red Sea/Gulf of Aden proto-ocean. 1 - locaUon of the Hadar Formation sequence shown In FIg. 7.

Effects of rifting on river drainage

MEDITERRANEAN SEA

s:::-}}}} Drll r n a 8~ n o' aHo" ~l ed L ... ,,:.:.::;.::.;. w il li rif! baul!

,./' I .uh ~, w ilh Ihr o w where known

Kil ome tr es

o 50

171

Flg. 5. RIver drainage and structure of the Gu1f of Suez ruft. 1, 2, 3 and 4 .. locations of the Abu Zentrna. Wadi Gharandal. Wadi Dara and Wadi Khar1m sedimentary sequences depicted In Flgs . 8 and 9:

g . Qa Plain: G .. Wadi Ghuwelbba: A .. Wadi Araba.

172 LYNNE FROSTICK and IAN REID

TABLE 1. Rift bastn catchments and the source of rivers.

Lake Total %Ax1al % Drainage % Drainage Basin Catchment Drainage Across Boundary Across Hanging

Area. kml faults WaH/Roil-over

Malawi 102,599 5.6 25.6 68.8

Tanganyika 151.685 9 .1 19.9 1.0

Turkana 148.403 57.5· 9 .9 32.6 (Present-

day)

Turkana 201.995 42.2- 7.3 SO.5 (9000 a

BP)

• the major part of this is the Orno river. a stream that flows for more than 90% orlts course outside of the rift

to the other in an apparently orderly manner (Rosendahl et al. 1986). Between the segments there are topographical highs (the transfer zones of Gibbs 1984: accommodation zones of Rosendahl etal. 1986) which ensure separation of the rift into a series of sub-basins. Each sub-baSin acts as a depocentre. even if the rill is flooded by marine incursion arby a lake (Cohen in press: Specht this volume).

At these early stages of rifling, integration of dratnage along the length of the rift is Inhibited by these structural and topographic highs. For exam­pIe , Lake Bogorta (990 m asJ) Is separated from the Lake Bartngo segment of the Gregory Rift by the Lobol Plain which lies only 9 m higher (Vincens et aL 1986). Because of this, true axial drainage Is rare. and It often contrtbutes little to the total catchment (see e.g. L.Tanganylka and L. Malawi, Table I) . Rivers captured by the developing basins and flOWing along the axis of the system often encounter a lake after comparatively short dis­tance . This Is the situation at the northern end of Lake Turkana. where the Omo River apparently acts as an axial sys tem. But It only spends its last 30 }an - < 4% of Its total length - within the rut (Butzer 1971). For the remaining 800 km of Its length it drains the Ethiopian highlands to the west of the rtft. and Is effectively backshed away from it. In fact, the Omo IS a good example of one rtft segment capturtng a rtver diverted away from another, adjacent. one. This happens frequently and results in the development of large fans or fan deltas at transfer zones.

Antithetic and Synthetic Faulting and Drainage

Small scale antithetic and synthetic faulting on the hanging wan block can produce slgnlfl.cant local diversions of drainage. ThIS is evident. for example, in the Culf of Suez (Fig. 5 , letter Q). Here, the rtver dra1ning the Qa Plain Is already within 12 km of the Sinal coast when It is caught behind a small antithellc fault and, as a result. runs parallel to the coast for over 40 km before entertng the Gulf. Slmllar diversions of drainage have been Inferred for ancient river systems draining across the hin­ged margin of a T'r1asslc half graben (Frostlck et at 19S5). An appreCiation of the possible influence of these smaner structural elements on sedimenta­tion has helped to provide a plaUSible explanation for an apparent conflict between different Indica­tors of palaeocurrent direction In a Situation where avaIlable outcrops are restrtcted in area.

Effect of Pre-existing Structures on Drainage Cross-cuttlng pre-rift structures can exert a

strong influence on the development of streams entering a r1ft bas in. and thererore on the localities or clastic sediment accumulation. The topographi­cal lows sometimes exploited by rivers that are attracted to a r1ft often correspond to faults or folds which pre-date the current phase or tectonism. Th1s IS the case on the northwestern shore of the Gulf of Suez., where a set or east-west trending faults probably generated durtng the emplacement or the Synan Arcs (Eocene; Chenet eta!. 1987) are exploited by the Wad lAraba and Wad! Ghuwelbba

Effects of rifting on river drainage 173

(Fig. 5. letters A and 0) on the hanging-wall block of the present M.

Other examples occur in the East African sys­tem. The River Lupa drains the Ruaha Rift (of Karroo agel. but now falls westward towards the Tertiary Rukwa R1fi at Its southern end. Further south. the RuUkrra rover. once part of an eastward draining system in the Ruhuha Rift (KarTOo) now drains westward to the northeast shores of Lake MalawI. There are yet other examples. especially in this part of East Africa where rtft systems of different age run almost orthogonal to each other.

Influence of Volcanism on Drainage The associal1on of volcaniSm with many rtfts.

especially during early phases of their develop­ment. exerts a strong local control on patterns of rtver drainage and sediment supply. Doming Is often a presage of a new volcanic centre. This In itselfw1ll divert drainage. but In extreme cases the volcanic pile has caused sub·divIslon of a basin. This Is the case In the Lakes Nakuru-Elmentelta­Nalvasha section of the Gregory RIft where the young volcanic centres of Longonot. Eburu and Menengal create topographic highs. Furthernorth. the rise of -rbe Barrler- little more than lO.CXX> years ago divided the1\J.rkana trough into 2 unequal parts rrruclde 1976: Dunkelman et aL this volume). The smaller. southern section • the Suguta - is now virtually staNed of water and as a result Is almost completely desiccated. Only a small salt lake - Lake Logipi - remains. fed largely by groundwater (Fig. 6). Stlll further north. Mount Kulal. a large shield volcano on the southeast shores of Lake Turkana. acts as a barner to rivers that might have drained westward from the KoroH and Chalbl deserts.

Later Stages of Rifting Recent Ideas concerning the evolution of contl­

nenlal rifts towards either paSSive marginS or failed rills (McKenzie 1978: Rosendahl 1987) pro­vide a starting POint for an understanding of the response of rtvers to later stages of rifling. As the system develops, much of the IniUal asymmetry is lost. Faults grow on both margins. It Is suggested that fault movement alternates through time from one side of the rtft to the other. There Is also general subSidence. In this structural setting, drainage across either margin may be Inhibited by back­tilting of fault blocks. These disruptions of the drainage network dlmlnish as fau It activity wanes In a failed rift or as faulling shifts offshore In a proto-ocean. But the general subsidence of the rtft axis has another. more Important. consequence for sedimentation. The sub-baSins that are a feature of the early stages of rifting are superceded by a continuous and linear topographic low along

the rift axis. This provtdes the opportUnity for development oflong axial rivers such as the Benue (Fig. 3). The sedimentary models of Brtdge and Leeder(1979) and LeederandAlexander(1987) are most approprtate for this stage of rtft development. These axial rtvers may be responsible for large deltaic wedges of matertal where a falled r1ft meets the ocean. as In the Niger Delta. However. If s ubSidence continues there may be marine inun­datton. In thiS case. the site of deltaic accumula· tlon will be controUed by the pOSItion of the shore­line.

Young passive margins often remain sediment staNed long after their initiation because the re­dlst:I1buUon of sediment by axtal rivers is preclu­ded by marine inundation. This Is the case v.1th the Red Sea system (Fig. 41 which receives no major river Inputs along the whole of Its 2000 kIn length.

PATTERNS OF SEDIMENTATION 1N RIFTS

The analysis of gross relationships between evolving structure and drainage together with recently developed facies models (Frostick and Reid 1987a: Leeder and Gawthorpe 1987) makes a good. starting pOint for an exam.1nation of sedimen­tary sequences In rifts. It would be advantageous to Identify features of sedimentation that are common despite the dJiferences in structural and environmental history that inevitably distinguish one system from another.

SeQ.uences in African Lake-fiUed Rifts Sediments in the East African rift have long

aroused interest (e.g. Reck 195 1; Isaac 1965: Pickford 1978: Frostickelai. 1986) . This Is largely because they contain an almost unique record of Cenozoic vertebrates Including our own early ancestors (e.g. Leakey and Leakey 1978). As a result. good stratigraphical and sedimentological infonnatlon Is readily available (see e.g. Bishop 1978; Frostick et al 1986). This provides a good basis for inter-basinal compartson of sedimentary sequences.

Examples of Plio-Pleistocene sequences from 3 dUferent lake basins are shown In Fig. 7. The sites are located on Figs. 2 and 4. It Is InterestJngtonote that these. and most of the other sequences re­ported in the literature. are for slles on the unfaul­ted margin of half graben I. e. on the roll-over or hanging-wail platform . This margin Is . topographically. the most subdued and the sequences are therefore likely to be complicated by repeated transgresSions and regreSSions as lake level rises and falls (Frosllck and Reid 1986). It is not. therefore. surprtslng that the 3 sequences depicted tn Fig. 7 consist of complex Interdlglta· lions oflake. lake·shore and fluviatile sediments.

174 LYNNE FRQSTICK and IAN REID

PRESENT DAY

,

M, " 'ow,, ,-to

• .•.. ," ,

9000a BP

L aka Turka"a d.alnag. bnln

Spillway

Vol".,.lc C . ... I •• Ifflcllng clra'".I11.1

Lake para""la" apllam.r.J

Fig. 6. Catchment of Lake Thrkana at present (1em and durtng the Holocene high lake stage 0[9000 a BP (Mght).

PedogeniC hortzons are common (e.g. Fig. 7, Kapthurtn Formation of Lake Baringo at approxJ­mately 20 ml. They represent pertods of relative quiescence in the landscape and may correspond With phases of boundary fault Inactivity. Eroded surfaces also occur sporadically throughout the sequence (e.g. Fig. 7 at 27 m in both the Hadar and Koobi Fora FonnaUonsl . These are related to

periods of river incision that might be caused by either tectonism or climatic change.

The most strUdng feature of Fig. 7 is the cycliCIty in all 3 sequences, each from Widely separated basins. Deposits of sut/ clay alternate with sand/ gravel. Each cycle varies in thickness between 0 .5 and 20 m . They reflect both the instab1l1ty of base­level and spasmodic changes in sedtment supply.

HADAR FORMATION

AFAR, ETHIOPIA

Alluvium

lake

KOOBI FORA FORMATION L. TURKANA,

KENYA

Alluvium

llake

Alluvium

~.... .J wll?Z2Z!f , .. " " ..

lake '."

Alluvium

lake

Alluvium

lake

Alluvium

KAPTHURIN FORMATION L. BARINGO,

KENYA

, ~

Pi .,-;": ,.tI

Key lake Margin ",; Slumps

-~ Uneven bedding

Alluvium 14)JI/ Cross-bedding

~4 Intrillc1ua

lake Margin .... Roots

.~ Shell dt,obris

""" Tuff

Alluvium - Gypulm

lake Margin T/fZ (onglomer.te

Safl duone

Ull Sihstone Alluvium 1- Claysto ne

Fig. 7 SedJmentary sequences from 3 East Afrtcan rift lake basins . Hadar Formation after TaJeb and nerce1Jn 1979; Koobi Fora. Formation after Bowen and Vondra 1973; Kapthurtn Formation after Tallon 1978. See FIgs. 2 and 4 for location.

rn ;;< a 0 -," " S"

'" 0 , ~ . • " 0-" • S" • ~

UI

176 LYNNE FROSTICK and I AN REID

Sequences in African Marlne Rifts Examples of Miocene to Pleistocene sequences

from the Gulf of Suez are given 1n Figs. 8 and 9. Their location is given in Fig. 5. The Abu Zenima section (Fig. 8) Is on the eastern shore close to the end of a major boundary fault. By contrast Wadi Gharandal (Fig.S) is further north and lies close in front of a boundary fault thought to have been acttve during Miocene Urnes. Wadi Dara and WadJ Khar1m (Fig. 9) are on the western margin of the

ABU ZENIMA

~~~~,." .. .... ,". Delta

" • • ... . ~ " pp'" •

Guifln an area where the present·day topography is subdued. At this point along the Gulf, the boundary fault is on the opposite shore and the sites are therefore on the hangmg-wall platfonn. This is a structural setting very s1m.1lar to that of the sites in the continental basins already de­scribed above.

Once more. all of the sequences exhlbUmarked cycllcity of sediment type and depositional envi­ronment. In these cases thIs largely reflects chan-

WADIGHARANDAL

. . . . .

>,. Alluvium

~ Shallow Marine ; , • ~

,

b.o.u 0 1"'" Beach

In" ", .. ,

.II' .... ~,. , ~

CI S,$dC p c 8

Shallow Marine

Beach

Alluvium

Basalt

• • • o o i • • ~

Marine Muds

Allu vium

(I S'Sd G P ( 8

Flg. B. Sedimentary sequences from the eastern coast of the Gulf of Suez. See Fig. 5 for location.

" c

" v 0 -, ~ "-

" c

" v o "-

10 m

5

o

Effects of rifti ng on river drainage 177

WADI DARA

. 0 · • 0 • · · 0

0 • . . • •

0 0 , · 0

• . . • 0 • 0 . • . ,

--fl~''.'')~~ .J .'~_~ ",":0

~ " I\... '-~,:" "~~ ~ '-" '\~~~~~ ~~,x..."',~

.. . · ,z,", .... ,: ...... ,', ...... ,', ...... If . ~ '-..0'-..''-'-''--= . . .... ' .

Kr~~'~ .»~'~,

. AI luvium

00 liths

B each luvium AI

B

AI

each

luvium

each luv ium

ea ch

B AI B

CI Si Sd C P CB

WADI KHARIM 15m

" " 0. 0. :0

10

5

o

Prode lta

o ' 0 . . (I Delta

Some biotu rba tion, large sc ale CfOSS bedding

o . ,

Prode lta

~ Delta 0

;/. . Some bioturbation,

. , large sca le cross bedding

II Delta Slumps

La on ..,....,..,....,~

CI Si Sd G PCB

Fig. 9. Sedimentary sequences from the western coast of the Gulf of Suez. See Fig. 5 for locaUon.

178 LYNNE FRQSTlCK and IAN REIO

ges in the relaUve levels of land and sea. The alternation of marine, marine shelf. beach and pro-delta sediments with more proximal delta and alluvial depoSits Is a feature of the Miocene to Recent sedimentation throughout the Gulf of Suez.

Frequently, the sequences contain thick evapo­lite deposits (Sellwood and Netherwood 19841. This is thought to reflect a sUuctura] lnstabUity that led repeatedly to the fonnaUon of a bamer at the mouth of the Gulf, cutting it off from the Mediterranean (Hassan and EI-Dashlouty 1970). However. evapOrites are a widely acknowledged feature ofmartne-inundated continental rift sedi­mentation (see e.g. Lowell and Genik 1972; Ponte etal. 1980) .

CAUSES OF CYCLICAL DEPOSITION IN RIFTS

Within an acUve rtft setting there is a temptation to explain most fluctuations in sedimentation In terms of the effects of faulting. There Is no doubt that movements on major boundary faults wUl have considerable long term effect on both the character and quantity of detritus supplied to the basin and on the relative levels of land and sea or lake (Blair 1987). By diverting rtvers. faulling may also cause abrupt switches in the location of clastIC sediment inputs . causing a local reduction in sedimentation rates and a change to non-clastic depos ition. But It Is frequently forgotten that other

Sea Level

changes in the character of the basin may also influence the style of sedimentation. Of special importance in lake basins are climatic fluctua­tions. whUe in marine-inundated rifts. a signifi­cant factor Is eustatic changes In sea-level.

c Um. tlc Change a nd Tectonism In Lake-filled Rifta .

Changes In the levels of the Afrtcan r1ft lakes on both short and long thue-scales are well documen­ted (Grove et al. 1975; Gasse etaL 1980: Reid and FrosUck 1986) . These changes are a response to a variety of factors that Interplay: seasonal ralnfall­e.g. the level of Lake Turkana changes by over a metre each year (Reid and Frostlck 1985); climatic change - e.g. Chew Bahlr in southern Ethiopia has all but drted out over the past 9000 years in response to Increased aridity (Grove et al. 1975): and volcano-tectonic alterations in basin shape -e.g. Lake Turkana Is thought to have been affected by the tumescence associated with Miocene lava (WatkJns 19861.

Two examples of African lake-level fluctuaUons over the past 40000 a. are shown in Fig. 10. These are for Lake Abhe in Ethiopia (Gasse 1977) and Lake Mobutu (formerly Albert) in Uganda (Adamson et al 1980), It Is difficult to be absolutely certain about the causes orth~5~ v;!.rtations tn lake level. Synchronous and sympathetiC changes in Widely separated bas ins tend to suggest that climatic change Is responsible. For example. between

African Rift Lake Levels I A~I ""HI I",K( MU"0I1U

,,- ' 0_ H'Ih

P I II '> I ()o r-~"----",= , ~ ~

(lN I >

• " " -. '. < > •

< " ..

< " .. ... -,

" " " 0 , ., L,

< " ,

" , 0

" -1 " 0'11.0 ' tiN'

Fig. 10. Eus1aUc sea-level curve (after VaU el at 1977) and lake level curves or2 African rift lakes (L. Abhe after Gasse 1977: L. Mobutu after Adamson et aL 1980).

EffeCIS of rifting on river drainage 179

10.000 and 9000 a. BP all of the Afrtcan lakes stood at a high level (Butzer etaL 1972: Adamson et at 1980: Owen et al. 19821. At this time It Is thought that rainfall was consIderably greater than at present. As a result of this particular change In water-balance. as well as at other Urnes (Frosllck & Reid 1980)' lake basins filled up and overflowed from one to the other.

Figure 6 shows the present Lake TUrkana catch­ment alongside an estimate of Us extent during the 9000 a. BP high lake stand. The catchment of the lake was enlarged considerably In the early Holocene by the oversp1ll of Chew Bahir and the Galla lakes through the Bakate Corridor. There was another signiflcanl, though smaller. exten­sion ofthe contrlbuUng drainage area In the south. Here. the Suguta (proto-Lake Logipi) overspilled westward through the Kamuge River to Join the Kerto. so circumventing The Bamer that had so recently erupted and divided U from Lake TUrkana (Truckle 1976). However. Lake Turkana Itself was now no longer the low poInt of a closed basin. It overflowed northwestward through the LogJpt depression to the Nile (Butzer et al1972; Hatvey and Grove 1982) . This high lake stand Is docu­mented by wtdespread diatomaceous silty clay and by reefs ofEtherla sp. almost 100m above the present lake level (FrosUck and ReId 1980; Owen and Renaut 1986).

A lack of coincidence between the lake level cutves for different basins does not necessartly rule out climatic change as a cause. Rainfall does not always increase or decrease by the same amount even over comparatively small distances. However. tectonJsm can only be invoked as a stgniflcant cause of lake level fluctuations if there is other corroborating evidence; e.g. the fine lake sediments which result from sudden basin subsi­dence being followed by an Influx of very coarse detritus from the hinterland (Blair 1987). or an IdentlficaUon of growth faults in the sedimentary record . Even wtthout tectoniSm, cyclical patterns of sedimentation such as those depicted in Fig. 7 can be produced. However, without the basin subsIdence that accompanies rifting. the great thicknesses of rhythmic sedimentary sequences that are seen in rifts could not be presetved.

Eustatic Changes ofSea-tevel and Tectonism in Mad ne-influence d Rifts.

Global curves of eustatic rise and fall In sea-level are wtdely used (Vall et al 1917: Fig. 10) despite resetvatlons about the methods employed in their construction. Although sea level Is not as unstable as lake level In internal drainage basins. fluctua­tions are both large and widespread. The succes­sive eustatic fluctuations of sea level wi1l be recor­ded as interbedded sequences of maline and al-

... Ul·f' .... .,

Juvial deposits. Many small-scale cycles of depOSition could be

linked to pertodiclty in sediment supply, wh Ich itself could have either a cUmatic or a tectonJc cause. The Wadi Gharandal sequence of the Gulf of Suez (Fig. 8) is the one presented here that is most Ukely to reflect the Influence offault acUvlty, since it is situated at the foot of what was, at the time of depOSition. a steep, acUve fault scarp. In this context it is interesting to note the relative SimpliCity ofthe sequence. wtth its repeated super­pOSition of coarse alluvium and fine marine muds and little evidence of reworking. This contrasts with the complexity of the Abu Zenima section. Which, although only 30 kIn along strtke. Is on the hanging-wall platform of the nelghbourl'1g half graben: the major boundary fau lt here is on the western side orthe Gulf. Changes in liver gradient brought about by fault aClivity have caused much reworking of sediment, while the lower topogra­phiC gradient has meant SignifIcant shuts of the shoreline during the transgreSSIons and regres­sions associated with eustatic changes In sea level. This difference In the complexIty of sedimen­tary sequences from one side of a half graben to the other is a feature of one of the latest facies models of rift basin sedimentation (Frostick and Reid 1987a) in which the most complex sequences occur on the unfaulted margin of the r1ft..

CONCLUSION

The topographical consequences of conUnental rifting are renected in the rtver system Which. In tum. controls the location, quantity and nature of clastic sediment supplied to r1ft basins. Doming in those rUts associated with mantle plumes combi­nes wtth backttltingoffootwaU fault blocks to shed rtvers away from the developing topographical lows. The faulted side of the basin becomes a by­pass margin and Is therefore. frequently sediment starved. For example. in Lakes Malawi. Tanganyt­ka and Turkana the drainage area of rtvers dJs­charging across boundary faults Is only 25. 20 and 10% respectJvely of the total catchment of each basin crable 1).

In East Africa and the Red Sea/Gulf of Aden. the catchments of the rift systems are surprtslngly small . Only a few. very small streams reach the Gulf of Aden. and only 3 medium-sized rivers - all ephemeral at present - discharge into the Red Sea along Us length of more than 4500 kro.

The same is true of the East African rtft. In some of the lake basins. e.g. Tanganyika. the total catch­ment Is only 4 times the surface area of the lake Itself. Major rtvers entering the rtft are few, and tend to uUlise either the accomodaUon zones bet­ween sub-basins where the obstrucUve boundary

180 LYNNE FROSTICK and I AN REID

faults die out. or the hanging-wall platform cum roll-over. The unfaulted margin ofthe rtft becomes the major source of sed1ment (Table 1). However, accommodation zones are frequently the sites of large-scale accumulations of clastic sedlments e.g. the Ketio RiveT delta on the western shore of Lake Turkana. In fact. the faulted margin may be an area of the basin that Is relatively starved of river sediment and may be a site of orgaruc-rich or carbonate deposits.

The fact that the rift Is a series of linked sub­basins during the early phases of development discourages the integrauon of drainage along the rift axis. However. in later phases. when the in­tensIty of faulting d1m1n1shes and regional subsi­dence plays an increastngly important role, large axial rivers may run the length of the rtft. as in the Senue trough. Whereas young rifts are compara­tively starved of sUiciclastic inputs and may accu­mulate organic-rich deposits which act as petro­leum source-rocks, later phases of rifting are characterised by widespread alluvium which may act as the potential reservoir rock.

The sedimentary fill of rift basins reflects not on1y the influence of tectonism, but also the effects of envirorunental changes that are independent of structural control. notably cUmatic fluctuations and eustatic changes In sea-level. In closed lake basins, shifts IncUmate produce changes In water­balance which are reflected In lake level . During wetter phases, lakes extend and may overllow one to the other. In marine basins , eustatic changes In sea-level will affect sedimentation In a way that is not dJsstmUar to the Influence of climatic change on lake basins. The rise and fall of sea-level wUl produce cyclical sedtrnentary sequences ahnost Identical to those produced as a consequence of boundary fault act1v1ty.

The lateral extent of the inundation resulting from transgression is of economic significance. Thts 1S because fine lake or marine sediments act as a seal to the more penneable alluvial deposits with which they interdigitate. The degree of Inun­daUon depends not only on the magnitude of the rise in water level. but also on topographic factors. In narrow. steep-sided rifts such as Tanganyika. a 5 m rtse in water level would shift the shoreline by 30 m (Evert 1980). In contrast. the same rise in water level in Lake Turkana would cause an average shift of 800 m (Butzer 1971).

The complex Interplay of tectonism. climatic change and eustatic adjustments of sea-level in rtft baSins makes 1t difficult to Isolate specific causes of changes in depositional style. There 1s a temptation to cUe tectonism as the major factor. However. this understates the role of the other factors and can lead to serious errors in palaeo­environmental reconstruction when COnSidering

ancient rift sediments. Acknowledgements- The authors' familiarity WIth the East Afrtcan rtfts artses from research financed by tile Natural Environment Research Councll of the UK. We would like to thank Richard Leakey for his help and encouragement over the years. The work In Egypt was made possible through a grant from Norsk Hydro pIc and by the generous assistance of the Egyptian Geological Survey. We are Indebted to All Mazzer for his help in the field. and to Bruce Sellwood for his advice before our fleldwork.

REFERENCES

Adamson. DA. Gasse. F .. Street, FA. & Williams. M.AJ. 1980. late Quaternary history of the Nile. Nature. 287. 50-5.

Allen. J.RL. 1965. Late Quaternary Niger Delta .. and adjacent areas: sedimentary environments and lithofacies. Bull. Am. Ass. Petrol. Geol .. 49.547-600.

Allen. J.RL. 1970. Sedimentsoflhe modem Niger delta: a summary and review. In: Morgan, J.P. (Ed). Deltaic SedfmentaUon. Mooem and Ancient Spec. Pub!. Soc. econ. Palaeont. Mineral. Tulsa. 15. 138-51.

Arambourg, C. 1933. Les formaUons pretertlares de la bordure occidentale de Lac Rodolphe (Afrtque Orientale). C.R. Acad. ScI. Paris, 197. 1663-5.

Baker. B.H. 1970. The structural pattern of the Afro­Arabian Rift system In relauon to plate tectonics. Phi.!. nans. R. Soc. London. 267. 383-91.

Baker. B.H. 19ti6. Tectonics and volcanism of the southern Kenyan Rift valley and Its Influence on rill sedlmentaUon. In: FrosUck. L.E. Renaut. RW .. Reid. I .. & TIercelin. J.J. (Eds). Sedimentation (n lheA.fri.can Rifts. Spec. Publ. Ceol. Soc. 25. 45-57.

Baker. B.H .. Williams. L.A.J .. Miller. TA & F'1tch. F.J. 1971. Sequence and geochronology of the Kenyan Rift volcanics. TecronophysicS, 11, 191 ·2 15.

Baker. B.H .. Williams, LAJ .. Miller. T.A.. & Mohr, PA 1972. Ceology of the Eastern Rift system of Afrtca. Spec. Pap. Grot Soc. Am.. 136. 67pp.

Bishop. W.W. 1978. Geological Background Ii) Fossil Man. ScotUsh Acad. Press. 585pp.

Blair. T.C. 1987. Tectonic and hydrologic controls on cycilc alluvial fan.fluvlal and Jacustr1ne r1fl-basln sedimentation. Jurassic-Lowermost Cretaceous Todos Santos FonnaUon. Chiapas. Mexico. Joumaloj Sedimentary Petroloqy. 1S7. 845-862.

Bowen. B.E .. & Vondra, C.F. 1973. StraUgraphlca] relationships of the Plio·Plelstocene depoSits. East Rudolf. Kenya. Nature 242. 391-393.

Br1dge. J.S .. & Leeder. M.R 1979. A simulation model of alluvial s traugraphy. Sedimentology. 26, 617-644.

Butzer. K.W. 1971.Recent hlstoryofan Ethiopian delta. Vniv. Chtcago. Res. Pap. Dept. Grog .. 136. 184pp.

Butzer. K.W.. Isaac. C.L.. Richardson. J.L. & Washboume-Kamau. C. 1972. Radiocarbon dating of East African lake levels. Scfence. 175, 1069-76.

Chapman. C.R. & Brook. M. 1978. Chronostratlgraphy of the Barlngo basin, Kenya. In: Bishop, W.w .. (Ed). Ceological Background to Fossil Man. ScotUsh. Acad. Press .. 207·24.

Chenet. P.Y .. Colletta. 8.. Letouzey. G .. Desforges. G ..

Effects of rifting on river drainage lSI

Ousset, E" & Zaghloul. E.A. 1987. Structures asso­ciated with extensional tectonics In the Suez rtft. In: Coward. M.P .• Dewey. J.F .. & Hancock. P.L. (Eds). Continental Extensional Tectonics. Spec. Pub!. Geol. Soc. 28. 551-8.

Cohen. A Facies relationships and sedimentation In large rift lakes: examples from lakes Turkana and Tanganyika. Palaeogeography, Palaeoclimatology. PolaeoecolDgy (In press).

Cohen. A & Thouln. C. 1987. carbonate sedimentation In Lake Tanganyika. Geology, 15,414-418.

CourUllot. V .. ArmiJO. R & Tapponnler, P. 1987, KinemaUcs of the Sinai triple Junction and a two· phase model of Arabla-Afr1ca rtflJng. In: Coward. M.P., Dewey, J.F. & Hancock. P.L. (Eds). Continental Extensjonal Tectonics. Spec. Pubi. Geol. Soc. , 28. 559-73.

Evert. M.J. 1980. Le Lac Tanganyika, sa Faune el la Peche au BuruncU. Unlv. Burundi Press. 200 p.

Falrhead. J.D. 1986. ~phySlca1 controls on sedimen­tation In the Afr1can Rift systems. In; Frostlck, L.E .. Renaut, RW .• Reid.!. &llercelln,J.J. (Eds), Sedtmen· lation In the Aft1can RIfls. Spec. Publ. Geol. Soc., 26. 19-27.

Fitch, F.J .. Hooker, P.J. & MlUer.J.A. 1978. Geochrono­logical problems and radioisotopic daUng In the Gregory Rift Valley. In: Bishop, W.W. (Ed). Geologbll Background (0 FOssU Man. ScotUsh. Acad. Press"

FrosUck. L.E., & Reid. I. 1980. Sorting mechanisms In coarse-grained alluvtal sedlmen ts: fresh evidence from a basalt plateau-gravel. Kenya. J .Ceol. Soc. LoncL, 137.431 -41.

FrosUck. L.E .. & Reld. I. 1986. Evolution and sedlmen· tary character of lake deltas fed by ephemeral r1vers In the 1\1rkana basin. northern Kenya. In: FrosUck, L.E., Renaut. RW .. Reid. I. & llercelln. J.J. Sedtmen· tatlon In the Aft1can Rifts. Spec. Publ. Geol. Soc. 25. 113·125.

Frostlck. L.E .. & Reid. I. 1987a. Tectonic controls of desert sediments In rtfl basins. ancient and modem. In: F'rostlck. L.E .• & Reid. 1. (Ecls). Desert Sediments: Ancient and Modem. Spec. Publ. Geol. Soc .• 35. 53-6B.

F'rosUck, L.E .. & Reid. 1. 1987b. A new look at r1rts. Geology Tcday 3 . 122-126.

FrosUck, L.E .. Reid. I .. Jarvis, J . & Eardley. H. 1988. n-tasslc sedlments of the Inner MorayF:lrth. Scotland: early r1fl deposits. J. Grol. Soc. Land .. 145. 235-248.

Frostlck. L.E .. Renaut. RW .. Reid. I .. & TIercelin. J.J. 1986. SedtmentatlDn In the Aji1can Rifts. Spec. Publ. Geol. Soc. 25. 382pp.

Garlunkel. Z .. & Bartov. Y. 1977. The tectonics of the Suez r1fl. BulL Geol. SUrv. Isr. 71. 1-44.

Gasse, F. 1977. EvoluUon oflakeAbhe (Ethiopia) from 70,000 B.P. Natw-e, 265. 42-5.

Gasse, F., Rognon. P .. & Street. FA 1980. Quaternary history of theNar and Ethiopian rift lakes. I n; Williams. M.AJ .. & Faure, H .. (Eds). The Sahara and (he Ntle. Balkema. Rotterdam. 361-400.

Gibbs. AD. 1984. Structural evoluUon of extensional basin margins. J. Geol.. Soc. Lond., 141. 609-20.

Grove. AT. 1986. Geomorphology of the African R.Ill system. In; FrosUck. L.E .• Renaut. RW .. Reid. I .. &

llercelln. J.J. (Eds). Sedtmentatfon In the Afiican Rift System. Spec. Pub!. Geol. Soc. 25. 9-16.

Grove. AT .. Street. FA. & Goudie, AS. 1975. Fonner lake levels and clImatlc change In the rill valley of southern Ethlopla.Geogr. J. 141. 177-202.

Harvey. P .• & Grove, AT. 1982. A prehlstortc source of the Nile. Geogr. J . 148. 372-36.

Hassan. F .. & EI-Dashlouty. S. 1970. Miocene evapo­rites of the GulfofSuez region and their Significance. Am. Assoc. Petrol. Geol. BulL 64. 1686-96.

Hubert. J .F .• Reed, AA & Carey, P.J. 1976. Palaeogeo· graphy of the East Berlin FormaUon. Newark Group. Connecticut Valley. Am. J. ScL 276. 1183-207.

Isaac, G.U. 1965. The stratlgTaphy of the PeninJ Beds and the provenance of the Natron AustraloplthlClne mandible. Quatemarfa. 7. 101·30.

Khan. A .. Maguire. P .. Henry. W .. & Higham. M. 1986. KRISP 85 - an Internatlonal seismic Invesugatlon of the Kenya run. Geology TOOay2. 139-44.

King, B.C. 1978. Structural and volcanic evolution of the Gregory R1ft valley. In: Bishop. W.W. (Ed) Geologi­cal Background to fbssU Man. Scottish. Acad. Press. 29-54.

Leakey. M.G. & Leakey. RE. 1978. Koobi Fora Research Project 1. 1he FossU Homlnlds and an introduction to their context. Clarendon Press, Oxford. 191pp.

Lefournler. J., ChorowtcZ, J .• Thouln, C .• Balzer. F .• Chenet, P .• Henrtet, J .. Masson. J., Mondeguer, A., Rosendahl. B .. Spy-Anderson, F .. & TIercelin. J.J. 1985. Le bassin de lac Tanganyika: evoluUon tecto­nique et5edlmentatrc:. Ac::ad. Scl CompteRend-, Parts 301. 1053-8.

Leeder. M.R. & Alexander. J . 1987. The or1gln and tectoniC significance of asymmetrical meander-belts. Sedimentology 34. 217·26.

Leeder. M.R. & Gawthorpe, RL. 1987. Sedimentary models forextenslonaJ tilt-block (half-graben) basins. In: Coward. M.P .. Dewey, J .F .• & Hancock. P.L. (Eds) Continental Extensional Tectonics. Spec. Publ. Geol. Soc. 28, 139-152

Lowell,J.D. &Genlk. G.J. 1972. Sea Ooorspreadlngand structural evoluUon of the southern Red Sea. Am. Assoc. Petrol. Geot Bull. 58. 247-57.

McKenzie, D.P. 1978. Some remarks on the develop­ment of sedlmentary basins. Earth Pfaner. ScL Letts. 40.25-32.

OJo. S.B .• & AJakalye. D.E. 1976. Prellmlnary Interpre­laUon of gravity measurements In the middle Niger Basin area. Nlger1a. In: Kogbe. C.A. (Ed), Geology oJ N!gena. Elizabethan Pub!. Co. 295·307.

Owen. RB .. Barthelme. J.W .. Renaut. RW. & Vincens, A. 1982. Palaeolimnology and archaeology of Holocene deposits north ·east of Lake 1\Jrkana. Kenya. Nature 298. 523-9.

Owen. RB .. & Renaut. R W. 1986. Sedimentology. stra­Ugraphy and palaeoenvironments of the Holocene Galana Bol FormaUon, north east Lake furkana. Kenya. In; Frostlck. L.E .. Renaut. RW .. Reid. I., & llercelln . J .J. (Eds) SedfmentaUon fn the AjrlcanRfjts. Spec. Publ. Geol. Soc 25. 311-22.

Pickford. M.H.L. 1978. Stratigraphy and mammalian palaeontology of the late Miocene Lukelno Fonnatlon. Kenya. In; Bishop. W.W. (Ed). GeologIcal Background

182 LYNNE FROSTtCK and IAN REID

to Fossil Man. Scottish. Acad. press., 268-7B. Ponte, F.L.. Fonseca. J.D., & Carozzl, AV. 1980. Petro­

leum habitats In the Mesozoic-Cenozoic of the conti­nental margin of Brazil. In: Mlall. AD. (Ed). Facts and Principles of World Petroleum Occurrence. Can. Soc. Petrol. Geol. Mem. 6, 857-66.

Reck, H. 1951. A preUmlnary surveyofthe tectonics and straUgraphy of OlduvaJ. In; Lt:akey. L.S.S. (Ed) Olduoof Gorge. 5-19.

Reid, I. . & Frostlck. L.E. 1985. Beach orientaUon. bar morphology and the concentration of metalliferous placer deposits: a case study. Lake lUrkana. J. OeoL Soc. LDnd 142, 837-48.

Reid, I., & FrosUck. L.E. 1986. Slope processes sediment derivaUon and landfonn evolution In a rift valley basin, northern Kenya. In: FrosUck. L.E., Renaut. R.W., Reid, I., &TIerceUn, J.J. (EdsJ Sedimentation in the Afr1can Rifts. Spec. Pub!. G~l. Soc. 25, 99-11l.

Rosendahl. B.R 1987. Architecture of continental rifts with speCial reference to East Africa. Ann. ReI). Earth. Planet. Sd. Letters 15, 445-503.

Rosendahl, B.R, Reynolds, D.J., Lorber. P.M., Bergess, C.F, McGW. J., Scott, D., Lambiase, J.J.,& Derksen, S.J. 1986. StnJctural expressions of rifting: lessons from Lake Tanganyika. Africa. In: FrostJck. L.E., Re­naut, RW., Reid. I., & TterceUn. J.J .. (Eels) Sedimen· tatlOn In the Ajrfcan Rifts. Spec. Publ. Geo!. Soc. 25. 29-43.

Sellwood, B.W., &Netherwood. RE. 1984. Facies evolu­tion in the Gulf of Suez area: sedimentation history as an Indlcalor or nft Inlt!aUon and development. Modem Geology 9, 43-69.

Taleb. M., & T1ercelln. J.J. 1979. SedImentation plio­cene et paleoenvtronnements de rift: exemple de la formation a Homlnldes d'Hadar (Afar. Ethlople). Bull. Soc. geol. France. 21. 243-53.

Tallon, P.W.J. 1978. Geological setting of the hominid fossils and Acheulian artifacts from the Kapthunn Formation, Bartngo District. Kenya. In: Bishop. W.W. (Ed). Geological Background to FossU Man. Scottish Acad. Press.361-73.

Truckle. P.H. 1976. Geology and the late Cenozoic sediments of the Suguta Trough. Kenya. Nature 263, 380-3.

Vall, P.R., Mitchum, RM., & Thompson, s. 1977. Seis­mic stratigraphy and global changes In sea-level, part four: global cycles of relaUve changes of sea-level. Am. Assoc. Petrol. Geo!. Mem. 26.

Vlncens. A, Casanova. J .. & T1ercelln, J.J. 1986. Palaeol1mnology of Lake Bogoria (Kenya) during the 4500 B.P. high lacustrine phase. In: FrostJck. L.E .. Renaut, RW" Reid. I .. & T1ercelin. J.J. (Eels). SedimentaUon in. the African Rifts. Spec. Publ. Geol. Soc. 25. 323-30.

Walkins. R T. 1986. Volcano-tectonic control of sedl­mentatJon In the Koobi Fora sedimentary basin. In: FrosUck. L.E .. Renaut, RW .. Reid, I., &T1erce1ln. J.J. (Eds). Sedimentation in. the Afr1can Rfjls. Spec. Pub!. Geol. Soc. 25, 85-95.

Williams. L.A.J .• & Chapman. G.R 1986. Relationships between major structures, sallc volcanism and sedi­mentation In the Kenya rift. from the equator northwards to Lake 1\J.rkana. In: Frostick, L.E., Renaut. RW., Reid. I .. & TIercel1n. J.J. (Eds). Sed[mentatlon in the J\frican Rifts. S~c. Publ. Geol. Soc. 25. 59-74.

WUlJamson. P.J" & savage. RJ.G. 1986. Early rtJl sedi mentation In theThrkanabasin. northern Kenya. In; Frostick. L.E., Renaut. RW., Reid, I., &T1ercelin, J.J. (Eels). Sedtmentatlon In the Afrtcan Rifts. Spec. Pub!. Geo!. Soc. 25. 267-83.