The drainage of Africa since Cretaceous

Abstract

Much of the drainage of Africa is relatively youthful. Many of its major rivers have shown substantial changes in their courses since the break up of Gondwanaland in the Cretaceous. In addition, many of the rivers have distinctive morphological characteristics such as inland deltas, cataracts and elbows of capture. Tectonic and climatic changes, including the development of the East African Rift System and the aridification of the Quaternary, help to explain the nature of these rivers. The history of the Saharan rivers, the Niger, the Nile, the Congo, the Cunene, the Zambezi, the Limpopo and the Orange, is reviewed.

Full text

The drainage of Africa since the Cretaceous
Andrew S. Goudie*
St. Cross College, Oxford University, Oxford, OX1 3LZ, UK
Received 27 June 2004; received in revised form 3 November 2004; accepted 3 November 2004
Available online 8 December 2004
Abstract
Much of the drainage of Africa is relatively youthful. Many of its major rivers have shown substantial changes in their
courses since the break up of Gondwanaland in the Cretaceous. In addition, many of the rivers have distinctive morphological
characteristics such as inland deltas, cataracts and elbows of capture. Tectonic and climatic changes, including the development
of the East African Rift System and the aridification of the Quaternary, help to explain the nature of these rivers. The history of
the Saharan rivers, the Niger, the Nile, the Congo, the Cunene, the Zambezi, the Limpopo and the Orange, is reviewed.
D2004 Elsevier B.V. All rights reserved.
Keywords: Africa; Drainage; Post-Cretaceous; Tectonics; Climate change
1. Introduction
Many of the rivers of Africa have a range of
intriguing characteristics that indicate that they have
had complex histories involving shifts in the nature
of their catchments and courses, particularly since
the splitting of Gondwanaland began in the early
Cretaceous. Nearly all of them are characterized by
falls (e.g. the Zambezi at Victoria Falls and Caborra
Bassa, and the Orange at Augrabies) and cataracts
(e.g. the Nile at Aswan, and the Congo at Stanley
Pool), some of them are linked by spillways (e.g. the
Zambezi and the northern Botswana drainage), many
of them drain into and out of large depressions (e.g.
the Nile out of the Sudd), some have inland deltas
(e.g. the Niger and the Okavango), some show
evidence of capture or piracy (as on the Cunene),
and some of them have become entangled in dune
fields (e.g. the Niger, Senegal and the Zambezi). In
addition, some drainage systems, such as those of
the northern Sahara or the mekgacha of the
Kalahari, have become defunct. These characteristics
reflect the long and varied climatic and tectonic
history of the continent and its distinctive topo-
graphic arrangement.
Unlike most other continents Africa is characterized
by passive rather than active plate margins and by a
dominance of basins, faults, rifts and topographic
swells rather than by compressional mountains
(Holmes, 1944; Summerfield, 1996; Doucoure´ and
de Wit, 2003). The Atlas Mountains are the only major
Cenozoic exception to this. There are, however, a
0169-555X/$ - see front matter D2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2004.11.008
* Tel.: +44 1865 271921; fax: +44 1865 278484.
E-mail address: Andrew.goudie@stx.ox.ac.uk.
Geomorphology 67 (2005) 437–456
www.elsevier.com/locate/geomorph

number of intraplate hotspots in Africa that have been
associated with Cenozoic doming and volcanism,
including those of Ahaggar, Tibesti, Jebel Marra and
Mt. Cameroon (King and Ritsema, 2000). These young
mantle plumes have been largely responsible for the
development of the basin and swell topography, which
is such a unique feature of the interior of the African
continent and which is so crucial in terms of drainage
development (Fig. 1a). In addition, the coastal margins
of much of Africa have been affected by a coastal
upwarp, which has its clearest manifestation in the
Great Escarpment of southern Africa. This marginal
upwarp accounts for the characteristic hypsometry of
Africa’s drainage basins, in which only a very small
proportion of the area of each basin is at low elevations
(Summerfield, 1991). Rifting has been another funda-
mental feature of the evolution of Africa, particularly
in East Africa (Fig. 1b).
The key events in the development of the setting of
African drainage basins have been:
(1) Opening of the South Atlantic (started ca. 150
Ma)
(2) Parana´-Etendeka volcanism (138–127 Ma)
(3) Final separation of Africa from S. America (ca.
100 Ma)
Fig. 1. (a) Some of the major plumes and basins of Africa. (b) The main rifts of East Africa.
A.S. Goudie / Geomorphology 67 (2005) 437–456438

(4) Start of eruption of Ethiopian and E. African
plateaus and arrival of Afar Plume head at
lithosphere (45 Ma)
(5) Start of E. African rifts and Red Sea rifting (30
Ma)
(6) Uplift of Red Sea margins begins (13.8 Ma) and
Gulf of Aden opens (10 Ma)
(7) Messinian salinity crisis and desiccation of the
Mediterranean (6 Ma)
(8) Start of Pleistocene aridity (ca. 2.4 Ma)
In this paper, the drainage histories of some of the
major basins are discussed and are related to the
tectonic and climatic vicissitudes of the continent.
What emerges from this is confirmation of the views
of early researchers such as Passarge (1904, p. 637)
and Falconer (1911, p. 241) that many of the river
courses of Africa are remarkably youthful. Fig. 2
shows the locations of the present major rivers of
Africa and some of the previous routes that have been
postulated for them.
Fig. 1 (continued).
A.S. Goudie / Geomorphology 67 (2005) 437–456 439

Some of the scenarios for drainage development
presented in this paper are speculative both with
regards to their dating and to their very existence,
but in recent years several developments have taken
place that have provided a greater degree of
certainty as to what has happened: remote sensing
has enabled certain ancient river courses to be
identified; radiometric dates have become available
for major phases of volcanic activity and rifting;
offshore cores, sonar and seismic studies have
indicated the fluvial source areas of depocentres;
tracer studies have been conducted in association
with mineral exploration (e.g. for diamonds); and
plate tectonic theory has provided a greater under-
standing of crustal instability.
2. North Saharan rivers
During the Neogene, uplift occurred in parts of the
Sahara (e.g. Tibesti and the Hoggar), and the climate
was sufficiently moist to enable large drainage
systems to develop. During the Miocene, four main
drainage basins were operational in the North Sahara:
the Eonile, the Eosahabi, the Gabes and the Chad
(Griffin, 2002). Of these, the Eosahabi and Gabes
Fig. 2. Major drainage lines of Africa at the present day together with some of the major former directions of drainage (shown by arrows)
discussed in this paper.
A.S. Goudie / Geomorphology 67 (2005) 437–456440

systems are now essentially inactive. The former
covered an area of around 0.9 million km
2
and drained
northwards into the Gulf of Sirt. The great diversity of
the fauna (including crocodiles) and the nature of the
flora suggest it was a large river which suffered
disruption and partial obliteration as a result of
increasing aridity in Pliocene and Pleistocene times.
The Gabes system was even larger, draining 1.1
million km
2
and entering the Gulf of Gabes. It passed
through what are currently closed depressions (e.g. the
Chott Melrhir and Chott el Jerid), but became
disrupted after about 4.6 Ma, when rainfall diminished
in the area. In addition, offshore seismic studies in the
Cape Bojador Slope also reveal that during the
Miocene substantial rivers flowed out of the Western
Sahara towards the Atlantic Ocean, depositing deltas
and eroding pre-existing sediments (Pickford, 2000).
In the eastern Sahara, a large system flowed from
Tibesti to Kufra (Pachur, 1993).
3. The Niger
The Niger (basin area 2.26 million km
2
)isa
remarkable river. It rises at an altitude of 800 m in the
Fouta Djallon plateau of Guinea, a mere 250 km east
of the Atlantic coastline in Sierra Leone and Liberia.
It then flows in a great arc (Fig. 3) that takes it over a
distance of 4200 km to reach the sea in Nigeria.
Initially, it flows towards the Sahara and near
Timbuktu passes through an inland delta. Then, at
Tosaye, it suddenly changes course and heads SSE
through Niger and into Nigeria. It is possible that the
Niger used to flow northeastwards into the Sahara,
creating a great lake in the Azaouad (Urvoy, 1942)
and that it has only recently flowed past the sill at
Tosaye. It is probable that its northward flow was
blocked by dune construction during arid phases of
the Pleistocene. In wetter phases, a lake formed
which in time spilled over the Tosaye sill to join the
lower Niger system (Tricart, 1965; Jacobberger,
1981; McIntosh, 1983). Large parts of the Niger
drainage catchment, as Fig. 3 shows, are currently
inactive, but during moist phases they were more
fully active and at high stages Lake Chad overflowed
over the Bongor Spillway into the Benue and thence
into the Niger (Talbot, 1980). During wet episodes,
drainage ran from the southwestern slopes of the
Ennedi massif into Lake Megachad and thence over
the spillway. This was a total distance of over 2500
km (Burke, 1996). It has also been argued, though
not universally accepted, that the Niger was in the
Tertiary fed by a major Trans-African palaeodrainage
system arising in Egypt and Sudan (McCauley et al.,
1982) and which has now been captured by the
growth of the Nile system. It is possible this phase of
augmented discharge could help to explain the
anomalously large size of the Niger delta fan. On
the other hand, the Niger delta, Africa’s largest, is
old. From ca. 80 to ca. 35 Ma, it was fed by a major
valley that occupied the site of the Benue Rift, which
had formed initially at ca. 140 Ma with the break up
of Gondwanaland.
4. The Nile
The Nile is another extraordinary river. It is nearly
7000 km long (and thus the longest river in the
world), drains some 3.2 million km
2
and stretches
approximately north to south over 358of latitude. It
manages to flow through one of the biggest tracts of
severe aridity on Earth, has numerous cataracts and
falls and yet has an immensely gentle gradient in its
lowest portion. Aswan, almost 1000 km from the sea,
lies at an altitude of only 93 m above present sea-
level. In spite of its great length and large catchment
area, its discharge is very small by the standards of
other rivers of its size. Moreover, as Said (1981, p. 6)
has pointed out, the Nile negotiates its way through
five regions which differ from one another in terms of
geological history and structure: the great Lake
Plateau of Central Africa, the Sudd and Central
Sudan, the Ethiopian Highlands, the Cataract tract
from Khartoum to Aswan, and the Egyptian region
down to the Delta and the Mediterranean.
The Egyptian Nile as we see it today is possibly a
young river in geological terms. It is only recently that
its diverse component tracts have been linked up and
that it has followed its present course through the
eastern Sahara. As Issawi and McCauley (1992) have
written:
Egypt, during the Cenozoic Era, was drained not by a
single master stream but by a succession of at least
three different, major drainage systems that competed
A.S. Goudie / Geomorphology 67 (2005) 437–456 441

Fig. 3. The maximum potential catchment areas of the Senegal and Niger–Benue systems. Areas that currently contribute to runoff are dotted. During periods of maximum humidity
(e.g. the early to mid-Holocene), the whole of the catchments may have been active. Overflow from the Chad Basin reached the Benue along the route indicated by the arrow
(modified after Goudie, 2002, Fig. 4.18).
A.S. Goudie / Geomorphology 67 (2005) 437–456442

for survival by means of gradient advantage. This
competition took place in response to tectonic uplifts
and sea-level changes during the interval between the
retreat of the Tethys Sea in late Eocene time (40 Ma
[million years ago]) and the birth of the modern Nile
during the late Pleistocene (~25 ka).
They present a possible model of Saharan Nile
evolution (Table 1). The first stage called the Gilf
system consisted of a northward flowing consequent
stream that followed the Tethys Sea as it retreated,
creating newly emergent land in Egypt. It also
consisted of some streams that formed on the flanks
of the Red Sea region towards the end of the Eocene.
The form of the Gilf system is shown in Fig. 4a. It is
probable that the climate during the Tertiary was
relatively humid and provided adequate run-off to
feed the system and to promote karst formation on the
Eocene limestones of the Western desert. This system
operated from ca. 40 to 24 million years ago.
The second stage, termed the Qena system, was
caused by major tectonic activity in the Red Sea area.
This caused a reversal of drainage to occur with a
large river flowing southwards towards Aswan and
the Sudan. The south-trending Wadi Qena is a relict of
this time and this helps to explain its curious direc-
tional geometry with respect to today’s Nile. The state
of the rivers during the Qena system times is shown in
Fig. 4b. It was at this time that catastrophic flooding
may have created some major erosional flutes and
gravel spreads in the Western Desert (Brookes, 2001).
The third stage, termed the Nile system (Fig. 4c),
was associated with base level changes in the
Mediterranean basin. In Messinian time, round about
6 million years ago, the Mediterranean dried up for
about 600,000 years (Krijgsman et al., 1999) because
of closure of the Straits of Gibraltar. Base level
dropped by 1000 m or more. Down cutting and
headward erosion became dominant processes and the
Eonile, as it is called, incised a deep gorge four times
larger than the Grand Canyon of the Colorado in the
USA. Because of its gradient advantage, it captured
the south-flowing Qena system and became the first
north-flowing river system that extended the length of
Egypt to the Mediterranean.
In the early Pliocene, sea-level rose to at least 125
m so that an estuary or ria extended more than 900 km
inland, reaching Aswan. This is termed the Gulf phase.
The subsequent stages of Nile evolution have been
explored by Said (1981). During the Paleonile phase a
local river occupied and filled the Gulf with sediments.
Following a dry phase, when the Nile ceased flowing,
the Prenile developed. This mid-Pleistocene event
appears to have been the first time that a mighty river
with a distant source reached Egypt. It drew some of
its discharge from Ethiopia when the Atbara (and
possibly the Blue Nile) pushed their way into Egypt
across the Nubian Swell by a series of cataracts. This
was caused by tilting in the Ethiopian Highlands so
that drainage was directed towards the Nile rather than
the Red Sea. Later still, during the Neonile phase, the
link with the Central Africa lake region was estab-
lished, but during the Pleistocene dry phases the Nile
was often seasonal rather than the perennial river it is
today. This was, for example, the case between ca.
20,000 and 12,500 years BP (Adamson et al., 1980;
Williams and Adamson, 1982).
Conversely, during more humid phases, like the
early Holocene, its flow was augmented by now
extinct tributaries such as the Wadi Howar, an 800-km
long watercourse that rose in the region between Jebel
Marra and Ennedi (Pachur and Kro¨pelin, 1987).
In the 1980s, there was speculation based on
analysis of remotely sensed radar data that a Trans-
Table 1
Issawi and McCauley’s stages of Saharan Nile evolution
1. Oldest—the Gilf system
Consists of north-flowing consequent streams that followed
the retreating Tethys Sea across the newly emerging lands of
Egypt and streams that formed on the flanks of the Red Sea
region towards the end of Eocene.
2. Middle—the Qena
Major south-flowing subsequent stream that developed along
the dip slope of zone of intensified uplift in the Red Sea
Range during the early Miocene. Flowed to Sudan basin.
Confined to west by retreating scarp of the Limestone Plateau
and on the east by the uplifted rocks of the Red Sea Range.
3. Youngest—the Nile system
Came into existence as a result of the drop in Mediterranean
sea-level in the late Miocene. Formerly local drainage eroded
headword into Limestone Plateau. Captured Qena system and
reversed its flow from south to north.
4. Pliocene flooding
After reopening of Straits of Gibraltar in early Pliocene
sea-level rose to at least 125 m. Estuary extended to Aswan
(N900 km inland).
5. Pleistocene sea-level change
(including Flandrian transgression)
A.S. Goudie / Geomorphology 67 (2005) 437–456 443

African drainage system, originating along the west-
ern margins of the Red Sea hills in Egypt, Sudan and
Ethiopia, found its way across to the Atlantic via Lake
Chad. (McCauley et al., 1982), though this was
criticized by Burke and Wells (1989) and has also
been challenged by Robinson et al. (2000), who
believed the so-called radar rivers of the Selima Sand
Sheet did not flow south westwards towards Chad but
northeast and eastnortheast towards Kharga.
Further towards it source, the Nile owes much of
its character to the uplift of the Ethiopian Highlands
and the development of the East African lakes. In
northwest Ethiopia, the volcanic plateau is a zone of
high relief that forms the headwaters of both the Blue
Nile and the Tekeze—the main tributaries (70% for
water and N95% for the suspended load) of the Nile.
The plateau results from regional basement uplift and
lava flow accumulation, associated with the develop-
ment of the Afar plume (Pik et al., 2003). The
volcanic activity was initiated in the Oligocene,
reaching a peak of activity at ca. 30 Ma ago. It is
also possible, on geochemical grounds, that there have
Fig. 4. (a) Sketch showing Gilf system (System I) at approximate end of Oligocene Epoch (Chattian Age, =24 Ma) (modified from Issawi and
McCauley, 1992). (b) Sketch showing Qena system (System II) in approximately middle Miocene time (Langian Age, about 16 Ma) (modified
from Issawi and McCauley, 1992). (c) Sketch showing Nile system (System III), which resulted from a major drop in sea-level of Mediterranean
in Messinian time (about 6 Ma) (modified from Issawi and McCauley, 1992).
A.S. Goudie / Geomorphology 67 (2005) 437–456444

been two plumes in East Africa, rather than one.
Rogers et al. (2000) have proposed that there was also
a Kenyan plume. On the other hand, Ebinger and
Sleep (1998) believed that Cenozoic magmatism
throughout East Africa could be the result of a single
plume.
In the Sudan, the Nile traverses the Sudd, which in
the Tertiary may have been a series of lake basins
(Salama, 1987), though whether such a lake existed
and how big it was is a matter of dispute (see Berry
and Whiteman, 1968).
The extent and catchments of the Central African
lakes which are related to the Nile system have varied
greatly according to rifting, tilting and climatic
changes. The present activity of the East African Rift
System started ca. 30 Ma ago (Burke, 1996) and by 25
Ma ago was active over a length of about 4000 km
stretching from the Gulf of Suez to the Mozambique
Channel. Subsidence and lake development has taken
place since that time. The Lake Malawi Basin, for
instance, has been actively subsiding since the late
Miocene (ca. 8.6 Ma) (Contreras et al., 2000).
Fig. 4 (continued).
A.S. Goudie / Geomorphology 67 (2005) 437–456 445

Neotectonic activity is still intense and associated
faulting has had a major influence on river courses
(Ve´ tel et al., 2004). Prior to the formation of the
Gregory Rift, about 15 Ma, East Africa seems to have
been dominated by a north–south trending arch of
crystalline rocks, which formed the continental water-
shed, with rivers draining east to the Indian Ocean and
west towards the Atlantic across the line of the present
Western Rift valley (Grove, 1983).
Lake Victoria, seems to be the result of regional
tilting and may only be 400,000 years old (Johnson et
al., 2000), and this also accounts for the form of Lake
Kyoga, the drainage reversal with which it is
associated (Fig. 5)(Doornkamp and Temple, 1966)
and for Pleistocene drainage evolution in the area
between Lake Victoria and Lake Edward in southern
Uganda. During Pleistocene dry phases some of the
lakes became closed saline systems and did not
overflow into the Nile. Lake Victoria only became
directly reconnected to the Nile system at ca. 13 ka
(Beuning et al., 2002). Conversely some lakes
currently cut off from the Nile (e.g. Chew Bahir and
Fig. 4 (continued).
A.S. Goudie / Geomorphology 67 (2005) 437–456446

Fig. 5. Changes in drainage pattern to the north and west of Lake Victoria. (a) The location of Lake Kyoga back tilted and flooded drainage. (b)
Major drainage lines before middle Pleistocene reversal (modified after Doornkamp and Temple, 1966). (c) Present drainage pattern.
A.S. Goudie / Geomorphology 67 (2005) 437–456 447

Turkana) overflowed into it during wetter phases via
the Sobat. Further south, Lake Rukwa has in the past
overflowed into Lake Tanganyika as a result of tiling
and Holocene climatic changes (Delvaux et al., 1998).
The date at which the complex sub-basins of the
Nile were linked up is the matter of ongoing debate.
Butzer and Hansen (1968, pp. 5–6) argued that as late
as mid-Tertiary times, prior to the major faulting that
produced the rift valleys of East Africa, the present Nile
basin was five or more separate hydrographic units.
Williams and Williams (1980, p. 221) suggested
that dthe first definitive evidence of Nile flow from
Ethiopia through Sudan to Egypt and the Mediterra-
nean is so far only of Lower Quaternary ageT, whereas
Adamson et al. (1993, p. 83) considered that dthe Nile
northwards from Ethiopia and from central to south-
ern Sudan dates from at least sometime in the
MioceneT. On the other hand, Berry and Whiteman
(1968) rejected the idea that the Nile is a very young
river and were of the opinion (p. 1) that:
...sections of the Nile system are very old and had
their origin in Late Cretaceous times or even earlier.
The date of connection is recent in the case of the
Bahr el Jebel and the Albert-Victoria sections, but in
the central and northern Sudan there must have been a
Nile valley in Late Cretaceous and Early Tertiary
times. Also in Ethiopia there must have been Miocene
and Pliocene Blue Nile systems.
5. The Zaire River
The Zaire (Congo) occupies a large basin in
central Africa, a basin which has a narrow exit to
the sea (Fig. 6). It has an area of 3.7 million km
2
,a
length of ca. 4100 km and an annual discharge of
ca. 1400 billion m
3
per year. Its tributaries create a
branching, leaf-like pattern that is unique in Africa.
Most of the major tributaries are oriented towards a
point in the centre of the basin rather than towards
its outlet (Summerfield, 1991, p. 425). It is also
notable for the falls that cross its course between
Kinshasha and the sea and for the fact that it has
no delta.
It seems possible that in the Tertiary (in the
Pliocene according to Beadle, 1974, p. 131) the
basin was occupied by a large water body (Peters
and O’Brien, 2001) of which lakes Tumba and
Maindombe are remnants. This internal drainage may
have been captured by a short stream draining to the
coast and cutting through the high ground at the
continental margin on the western rim of the basin.
The point of capture was possibly just below Stanley
Pool between Brazzaville and Kinshasha. Burke
(1996) has suggested that this may have happened
in the last 30 Ma. Terrigenous sediment supply to the
West African margin increased dramatically in the
Miocene (Lavier et al., 2001) and it is conceivable
that this is when the Congo System was firmly
linked to the Atlantic. In any event, doming
associated with rifting in East Africa in the Miocene
would have augmented its flow (Uenzelmann-Nebel,
1998). During the upper Cretaceous, the Ogooe and
Kwanza fans had been the most important loci of
sedimentation along the West Africa margin, as was
the Kouilou/Niari river (Uenzelmann-Nebel, 1998).
Lucazeau et al. (2003; Leturmy et al., 2003) surmise
that the switch of depocentres and the capturing of
the endoreic Congo system may be associated with
the flexural uplift created by sediment loading on the
Continental Shelf.
6. The Cunene and Coroca
The Cunene rises in the Bie´ highlands of Angola
(Fig. 7) and flows into the Atlantic on the border
between Angola and Namibia. For much of its course,
it flows southwards, as if towards the Etosha Pan, an
ancient structural basin (Buch and Trippner, 1997),
but then it turns sharply westwards and enters a tract
with steep falls and rapids (e.g. the Caxambue rapids
and Ruacana Falls). This seems to be the case of a
river capture by a stream eroding backwards from the
coast and capturing the interior drainage (Wellington,
1955, p. 65). Wellington also suggests that the
conditions for a similar process of capture are present
in the headwaters of the Rio Coroca to the north of
the lower Cunene. This river, having eroded head-
ward through the Sierra de Chela of the Great
Escarpment, is threatening to behead the upper
Caculuvar River, a tributary of the upper Cunene.
The timing of the Cunene capture is not well
constrained (Moore and Blenkinsop, 2002) but low-
ering of the base-level associated with the opening of
A.S. Goudie / Geomorphology 67 (2005) 437–456448

the Atlantic may have initiated a period of rapid
erosion, which may have exploited a Perma–Carbon-
iferous glacial valley from which Karoo sediments
were stripped (Martin, 1950). It has also been argued
that the capture, which perhaps caused the shrinkage
of a postulated large Lake Etosha, occurred only ca.
35,000 years ago (Buch, 1997). At the far western end
of its course, as it passes through the coastal sand sea,
there are signs that the Cunene formerly entered the
sea considerably to the south of its present mouth
(Sander, 2002) and that it may have been forced
northwards by dune encroachment.
7. The Zambezi
Relatively abrupt changes in the direction and
characteristics of its modern course suggests that the
Zambezi (which covers an area of ca. 1.33 million
km
2
) may only relatively recently have become a
joined up system. The upper and middle Zambezi are
thought to have evolved as separate systems, with
the upper Zambezi previously joined to the Limpopo
system and the middle Zambezi to the Shire system.
It has been argued that the upper Zambezi was
captured because of down-warping and tectonic
Fig. 6. The Congo (Zaire) Basin. The 500-m contour emphasizes the approximate boundaries of the relatively flat central basin, which may have
formerly contained a large lake or lakes.
A.S. Goudie / Geomorphology 67 (2005) 437–456 449

enhancement of the stream power of the Palaeo–
middle Zambezi (Shaw and Thomas, 1988, 1992).
This linking of the two systems may have been as
recent as the Pliocene or mid-Pleistocene (Thomas
and Shaw, 1991, p. 34). The proposed nature of the
drainage shifts involving the Zambezi, Kafue,
Luangwa and Shire are shown in Fig. 8. After the
establishment of the present Zambezi course,
enhanced downwarping along the Gwembe trough
caused rejuvenation resulting in the rapid develop-
ment of the Victoria Falls and the formation of the
incised middle Zambezi gorge.
8. The Orange River
TheOrangeRiver,withacatchmentareaof
953,200 km
2
, is the largest basin in southern Africa.
It has shown considerable changes in its pattern and
extent since the Cretaceous (Dollar, 1998), and these
have been the subject of considerable recent research
because old river courses may be sources of diamonds
(Moore and Moore, 2004; Jacob et al., 1999).
During the Oligocene, offshore sediment studies
indicate that the upper Orange–Vaal system entered
the South Atlantic through the Cape Canyon, some
300–500 km south of its present mouth at Orange-
mund. Subsequently, towards the end of the Miocene,
river capture by the Koa River and its tributary
headwaters resulted in the diversion of the upper
Orange to its present course. The Koa Valley
contained a north-flowing perennial system during
the middle Miocene. Even later, further tectonic
activity and aridification (De Wit, 1999)caused
abandonment of the flow through the Koa (Dingle
and Hendey, 1984). The Orange may also have
drained a large area of the Kalahari, being fed by
the Trans-Tswana river of McCarthy (1983). The
Molopo, Morokweng and Harts were major north
bank tributaries (Bootsman, 1997; Bootsman et al.,
1999). Tectonism along the Griqualand–Transvaal
Axis and the associated development of the Kalahari
basin caused the demise of these important palaeo-
tributaries (Moore, 1999). Just how extensive the
Trans-Tswana river was, however, is a matter of
contention (Fig. 9), for although McCarthy (1983)
Fig. 7. The Cunene and Coroca rivers.
A.S. Goudie / Geomorphology 67 (2005) 437–456450

Fig. 8. Drainage development in southern Africa. (a) Major modern drainage lines. (b) The drainage system postdating the division of
Gondwanaland. (c) The situation prior to the union of the middle and upper Zambezi in the early Pleistocene (modified after Thomas and Shaw,
1991, Fig. 2.8).
A.S. Goudie / Geomorphology 67 (2005) 437–456 451

Fig. 9. The hypothetical Trans-Tswana drainage (modified from Dardis et al., 1988, Fig. 2.8).
Fig. 10. Reconstruction of the palaeodrainage of the Orange River drainage system. (a) Late Cretaceous, (b) Palaeogene, (c) Neogene (modified
after Dingle and Hendey, 1984).
A.S. Goudie / Geomorphology 67 (2005) 437–456452

suggested that the catchment area may have extended
into tropical Central Africa, and taken in the Kafue,
upper Zambezi and possibly the Kwando and Oka-
vango, De Wit et al. (2000) suggest that at least part of
that area was drained by the palaeo-Limpopo River.
De Wit (1999) has argued that during the late to
middle Cretaceous there were two main rivers
draining the interior of Southern Africa. The south-
ern one, termed the Karoo River, had its source in
the present Orange/Vaal basin, but had its outlet at
the present Olifants River mouth, a finding supported
by biogeographical studies (Barber-James, 2003).
The northerly one, termed the Kalahari River,
drained southern Botswana and Namibia and entered
the Atlantic via the lower Orange River (Fig. 10). By
the early Cenozoic, however, the lower Kalahari
River had captured the upper part of the Karoo
River. This he attributed to accelerated uplift of the
southern and eastern subcontinental margins at
around 100–80 million years ago. The position of
the Atlantic outlet of the Kalahari River is disputed,
however, and Stevenson and McMillan (2004)
suggest that in the Latest Cretaceous the dominant
exit was situated offshore the present-day Groen to
Buffels rivers.
In Southern Africa, it is possible that the
drainage pattern was affected by doming over large
mantle plumes (radii of ca. 1000 km) that
succeeded continental breakup in the Cretaceous.
In the west there was the so-called Parana´ plume
and in the east the earlier so-called Karoo plume
(Moore and Blenkinsop, 2002). At least in some
areas, drainage radiated away from the centre of the
plumes (Cox, 1989)(Fig. 11), but the precise
location and number of the plumes is the matter
of debate. Some flank drainages that developed on
the Parana´ plume could include the Fish and
Molopo rivers, the Okavango, Cuango and upper
Zambezi and Congo headwaters. This could also
explain the striking absence of major westward
draining systems between the modern Orange and
the Congo.
Fig. 11. Postulated locations of major domes associated with Karoo and Parana´ mantle plumes and associated drainage patterns (modified after
Cox, 1989).
A.S. Goudie / Geomorphology 67 (2005) 437–456 453

In addition, as elsewhere in Africa (e.g. the Congo,
Sudd and Chad basins), interior basins subsided to
accommodate sedimentary material in the Cenozoic
and these were bounded by uplifting sub-swells such
as the Transvaal–Griqualand axis and the Okavango–
Kalahari–Zimbabwe axis (Moore and Larkin, 2001).
Uplift of these sub-swells contributed to drainage
dismemberment (Gumbricht et al., 2001; Partridge
and Maud, 1987) and, for example, severed the links
between the Limpopo and the Okavango, Cuando and
Zambezi-Luangwa system (Moore and Larkin, 2001).
The Great Escarpment on the Atlantic coast was also
affected by flexural isostatic uplift, with denudation
contributing to rift-margin upwarping (Gilchrist and
Summerfield, 1990).
In the heart of the Kalahari sandveld, there are a
number of defunct drainage systems, mekgacha,
which appear to have been active during Quaternary
pluvials but which are now unable to flow (Shaw et
al., 1992).
9. Dune aligned drainage
Recent analysis of remotely sensed imagery has
revealed the widespread existence of aligned drain-
age systems associated with past aeolian activity and
dune development across much of Africa. The
former expansion of the Sahara southwards towards
the Gulf of Guinea and of the Kalahari northwards
towards the Congo Basin mean that major dune
fields developed which guided the pattern of river
courses in subsequent more humid phase. Aligned
drainage of this type has been described from
southern Angola (Shaw and Goudie, 2002), the
Democratic Republic of Congo (de Dapper, 1988),
and in southern Nigeria and Cameroon (Nichol,
1998, 1999). There is also an extensive area of
aligned drainage with karstic development in the
Weissrand Plateau area of southeastern Namibia. It is
developed on a Tertiary calcrete surface from which
the dunes are now largely absent.
10. Conclusions
In common with Australia (Beard, 2003) and South
America (Potter, 1997), it is clear that since the break
up of Gondwanaland, the drainage systems of Africa
have shown substantial changes. Prior to rifting,
drainage may have developed from mantle plumes.
After that, there was uplift of the bounding mountains
(e.g. the Great Escarpment of southern Africa),
subsidence of internal basins (e.g. the Kalahari) and
uplift of subswells produced by local mantle upwell-
ing (Burke, 1996). Aggressive rivers, adapting to a
new low base level, cut back from the coast and a
suite of river captures occurred (e.g. the Cunene and
the Congo). In eastern Africa, the growth of the Afar
plume, the building of the Ethiopian Plateau and the
Red Sea Hills, and the development of the rift valleys
caused further fundamental changes in river catch-
ments. In North Africa, the retreat of the Tethys Ocean
and the Miocene desiccation of the Mediterranean
(with its associated low base level) had a major impact
on the Nile and other systems (Griffin, 2002).
Accentuated aridification in the Pleistocene caused
dune fields to expand and lake basins to shrink and
close, created areas of aligned drainage (as in South
Angola), blocked and diverted drainage (e.g. the
Niger), and isolated lakes from major river systems
(e.g. Lakes Victoria and Rukwa).
References
Adamson, D.A., Gasse, F., Street, F.A., Williams, M.A.J., 1980.
Late Quaternary history of the Nile. Nature 287, 50 – 55.
Adamson, D., McEvedy, R., Williams, M.A.J., 1993. Tectonic
inheritance in the Nile basin and adjacent areas. Isr. J. Earth-Sci.
41, 75– 85.
Barber-James, H.M., 2003. The biogeography of the Prosopisto-
matidae with a particular emphasis on southern African species.
In: Gaino, E. (Ed.), Research Update on Ephemeroptera and
Plecoptera. University of Perugia, Perugia, pp. 263 – 270.
Beadle, L.C., 1974. The Inland Waters of Tropical Africa. Long-
man, London.
Beard, J.S., 2003. Paleodrainage and the geomorphic evolution of
passive margins in southwestern Australia. Z. Geomorphol. 47,
273 – 288.
Berry, L., Whiteman, A.J., 1968. The Nile in the Sudan. Geogr. J.
134, 1– 33.
Beuning, K.R.M., Kelts, K., Russell, J., Wolfe, B.B., 2002.
Reassessment of Lake Victoria–Upper Nile River paleohydrol-
ogy from oxygen isotope records of lake sediment cellulose.
Geology 30, 559 – 562.
Bootsman, C.S., 1997. Indicators of the evolution of the upper
Molopo drainage. S. Afr. Geogr. J. 79, 83 – 92.
Bootsman, C.S., Reimold, W.V., Brandt, D., 1999. Evolution of the
Molopo drainage and its possible disruption by the Morokweng
A.S. Goudie / Geomorphology 67 (2005) 437–456454

impact event at the Jurassic–Cretaceous boundary. J. Afr. Earth
Sci. 29, 668 – 678.
Brookes, I.A., 2001. Possible Miocene catastrophic flooding in
Egypt’s western desert. J. Afr. Earth Sci. 32, 325– 333.
Buch, M.W., 1997. Etosha pan—the third largest lake in the world?
Madoqua 20, 49 – 64.
Buch, M.W., Trippner, C., 1997. Overview of the geological and
geomorphological evolution of the Etosha region, northern
Namibia. Madoqua 20, 65 – 74.
Burke, K., 1996. The African plate. S. Afr. J. Geol. 99, 341 – 409.
Burke, K., Wells, G.L., 1989. Trans-African drainage system of the
Sahara: was it the Nile? Geology 17, 743– 747.
Butzer, K.W., Hansen, C.L., 1968. Desert and River in Nubia.
University of Wisconsin Press, Madison.
Contreras, J., Anders, M.H., Scholz, C.H., 2000. Growth of a
normal fault system: observations from the Lake Malawi basin
of the East African rift. J. Struct. Geol. 22, 159 – 168.
Cox, K.G., 1989. The role of mantle plumes in the development of
continental drainage patterns. Nature 242, 873 – 877.
Dardis, G.F., Beckedahl, H.R., Stone, A.W., 1988. Fluvial
systems. In: Moon, B.P., Dardis, G.F. (Eds.), The Geomorphol-
ogy of Southern Africa. Southern Book Publishers, Johannes-
burg, pp. 30 – 56.
de Dapper, M., 1988. Geomorphology of the sand-covered plateaux
in southern Shaba, Zaire. In: Dardis, G.F., Moon, B.P. (Eds.),
Geomorphological Studies in Southern Africa. Balkema, Rot-
terdam, pp. 115 – 135.
Delvaux, D., Kervyn, F., Vittori, E., Kajara, R.S.A., Kilembe, E.,
1998. Late quaternary tectonic activity and lake level change in
the Rukwa Rift Basin. J. Afr. Earth Sci. 26, 397– 421.
De Wit, M.C.J., 1999. Post-Gondwana drainage and the develop-
ment of diamond placers in western south Africa. Econ. Geol.
94, 721– 740.
De Wit, M.J.C., Marshall, T.R., Partridge, T.C., 2000. Fluvial
deposits and drainage evolution. In: Partridge, T.C., Maud, R.R.
(Eds.), The Cenozoic of Southern Africa. Oxford University
Press, New York, pp. 55 – 72.
Dingle, R.V., Hendey, Q.B., 1984. Mesozoic and Tertiary sediment
supply to the western Cape Basin and paleodrainage systems in
south-western Africa. Mar. Geol. 56, 13 – 26.
Dollar, E.S.J., 1998. Palaeofluvial geomorphology in southern
Africa: a review. Prog. Phys. Geogr. 22, 325 – 349.
Doornkamp, J.C., Temple, P.H., 1966. Surface, drainage and tectonic
instability in part of southern Uganda. Geogr. J. 132, 238 – 252.
Doucoure´, C.M., de Wit, M.J., 2003. Old inherited origin for the
present near-bimodal topography of Africa. J. Afr. Earth Sci. 36,
371 – 388.
Falconer, J.D., 1911. The Geology and Geography of Northern
Nigeria. Macmillan, London.
Ebinger, C.J., Sleep, N.H., 1998. Cenozoic magmatism throughout
east Africa resulting from impact of a single plume. Nature 395,
788 – 791.
Gilchrist, A.R., Summerfield, M.A., 1990. Differential denudation
and flexural isostasy in formation of rifted-margin upwarps.
Nature 346, 739 – 742.
Goudie, A.S., 2002. Great Warm Deserts of the World: Landscapes
and Evolution. Oxford University Press, Oxford.
Griffin, D.L., 2002. Aridity and humidity: north Africa and the
Mediterranean. Palaeogeogr. Palaeoclimatol. Palaeoecol. 182,
65 – 91.
Grove, A.T., 1983. Evolution of the physical geography of the east
African Rift Valley Region. In: Sims, R.W., Price, J.H.,
Whalley, P.E.S. (Eds.), Evolution, Time and Space: the
Emergence of the Biosphere. Academic Press, London and
New York, pp. 115 – 155.
Gumbricht, T., McCarthy, T.S., Merry, C.L., 2001. The topography
of the Okavango Delta, Botswana, and its tectonic and
sedimentological implications. S. Afr. J. Geol. 104, 243 – 264.
Holmes, A., 1944. Principles of Physical Geology. Thomas Nelson
and Sons, Edinburgh.
Issawi, B., McCauley, J.F., 1992. The Cenozoic rivers of Egypt: the
Nile problem. In: Adams, B., Friedman, R. (Eds.), The
Followers of Horus. Oxbow Press, Oxford, pp. 1 – 18.
Jacob, R.J., Bluck, B.J., Ward, J.D., 1999. Tertiary-age diamond-
iferous fluvial deposits of the Lower Orange River Valley,
southwestern Africa. Econ. Geol. 94, 749 – 758.
Jacobberger, P.A., 1981. Geomorphology of the upper inland Niger
delta. J. Arid Environ. 13, 95 – 112.
Johnson, T.C., Kelts, K., Odada, E., 2000. The Holocene history of
Lake Victoria. Ambio 29, 2 – 11.
King, S.D., Ritsema, J., 2000. African hot spot volcanism: small-
scale convection in the upper mantle beneath cratons. Science
290, 1137 – 1140.
Krijgsman, W., Hilgen, F.J., Raffi, I., Sierro, F.J., Wilson, D.S.,
1999. Chronology, causes and progression of the Messinian
salinity crisis. Nature 400, 652 – 655.
Lavier, L.L., Steckler, M.S., Brigaud, F., 2001. Climatic and
tectonic control of the Cenozoic evolution of the west African
margin. Mar. Geol. 178, 63 – 80.
Leturmy, P., Lucazeau, F., Brigaud, F., 2003. Dynamic interactions
between the Gulf of Guinea passive margin and the Congo River
drainage basin: 1. Morphology and mass balance. J. Geophys.
Res. 108 (B8).
Lucazeau, F., Brigaud, F., Leturmy, P., 2003. Dynamic
interactions between the Gulf of Guinea passive margin
and the Congo River drainage basin: 2. Ioostasy and uplift.
J. Geophys. Res. 108 (B8).
Martin, H., 1950. Sqdwestafrika. Geol. Rundsch. 38, 6 – 14.
McCarthy, T.S., 1983. Evidence for the former existence of a major,
south flowing river in Griqualand west. Trans. - Geol. Soc.
South Afr. 86, 37 – 49.
McCauley, J.F., Schaber, G.G., Breed, C.S., Grolier, M.J., Haynes,
C.V., Issawi, B., Elachi, C., Blom, R., 1982. Subsurface valleys
and geoarchaeology of the eastern Sahara revealed by shuttle
radar. Science 218, 1004 – 1020.
McIntosh, R.J., 1983. Floodplain geomorphology and human
occupation of the upper inland delta of the Niger. Geogr. J.
149, 182 – 201.
Moore, A., 1999. A reappraisal of epeirogenic flexure axes in
southern Africa. S. Afr. J. Geol. 102, 363 – 376.
Moore, A., Blenkinsop, T., 2002. The role of mantle plumes in
the development of continental-scale drainage patterns: the
southern African example revisited. S. Afr. J. Geol. 105,
353 – 360.
A.S. Goudie / Geomorphology 67 (2005) 437–456 455

Moore, A., Larkin, P.A., 2001. Drainage evolution in south-central
Africa since the break up of Gondwana. S. Afr. J. Geol. 104,
47 – 68.
Moore, J.M., Moore, A.E., 2004. The roles of primary kimberlitic
and secondary Dwyka glacial sources in the development of
alluvial and marine diamond deposits in southern Africa. J. Afr.
Earth Sci. 38, 115 – 134.
Nichol, J.E., 1998. Quaternary climate and landscape development
in west Africa: evidence from satellite images. Z. Geomorphol.
42, 229 – 347.
Nichol, J.E., 1999. Geomorphological evidence and pleistocene
refugia in Africa. Geogr. J. 165, 79 – 89.
Pachur, H.-J., 1993. Palaeodrainage systeme im Sirte-Becken und
seiner Umrahmung. Wurzbg. Geogr. Abh. 87, 17 – 34.
Pachur, H.-J., Kro¨pelin, S., 1987. Wadi Howar: paleoclimatic
evidence from an extinct river system in the southwestern
Sahara. Science 237, 298 – 300.
Partridge, T.C., Maud, R.R., 1987. Geomorphic evolution of
southern Africa since the Mesozoic. S. Afr. J. Geol. 90,
179 – 208.
Passarge, S., 1904. Die Kalahari. D. Reimer, Berlin.
Peters, C.R., O’Brien, E.M., 2001. Palaeo-lake Congo: implications
for Africa’s late Cenozoic climate—some unanswered ques-
tions. Palaeoecol. Afr. 27, 11 – 18.
Pickford, M., 2000. Crocodiles from the Beglia formation, middle/
late Miocene boundary, Tunisia, and their significance for
Saharan palaeoclimatology. Ann. Pale´ ontol. 86, 59 – 67.
Pik, R., Marty, B., Carignan, J., Lave´, J., 2003. Stability of the
upper Nile drainage network (Ethiopia) deduced from (U–Th)
thermachronometry: implications for uplift and erosion of the
Afar plume dome. Earth Planet. Sci. Lett. 215, 73 – 88.
Potter, P.E., 1997. The Mesozoic and Cenozoic paleodrainage of
South America: a natural history. J. South Am. Earth Sci. 10,
331 – 344.
Robinson, C., El-Baz, F., Ozdogan, M., Ledwith, M., Blanco, D.,
Oakley, S., Inzana, J., 2000. Use of radar data to delineate
paleodrainage flow directions in the Selima Sand Sheet.
Photogramm. Eng. Remote Sensing 66, 745 – 754.
Rogers, N., Macdonald, R., Fitton, J.G., George, R., Smith, M.,
Barreiro, B., 2000. Two mantle plumes beneath the East African
Rift System: Sr, Nd, and Pb isotope evidence from Kenya Rift
basalts. Earth Planet. Sci. Lett. 176, 387 – 400.
Said, R., 1981. The Geological Evolution of the River Nile.
Springer Verlag, New York.
Salama, R.B., 1987. The evolution of the River Nile. The buried
saline rift lakes in Sudan el Arab Rift, the Sudd buried saline
lake. J. Afr. Earth Sci. 6, 899 – 913.
Sander, H., 2002. Relief- und Regolithgenese im nordo¨stlichen
Kaokoland/Namibia. Doctoral thesis, University of Ko¨ln.
Shaw, A., Goudie, A.S., 2002. Geomorphological evidence for the
extension of the Mega-Kalahari into south-central Angola. S.
Afr. Geogr. J. 84, 182 – 194.
Shaw, P., Thomas, D.S.G., 1988. Lake Caprivi: a late Quaternary
link between the Zambezi and middle Kalahari drainage.
Z. Geomorphol. 32, 329 – 337.
Shaw, P., Thomas, D.S.G., 1992. Geomorphology, sedimentation
and tectonics in the Kalahari Rift. Isr. J. Earth-Sci. 41, 87 – 94.
Shaw, P.A., Thomas, D.S.G., Nash, D.J., 1992. Quaternary fluvial
activity in the dry valleys (mekgacha) of the middle and
southern Kalahari, southern Africa. J. Quat. Sci. 7, 273 – 281.
Stevenson, I.R., McMillan, I.K., 2004. Incised valley fill stratigraphy
of the upper cretaceous succession, proximal Orange Basin,
Atlantic margin of southern Africa. J. Geol. Soc. 161, 185 – 208.
Summerfield, M.A., 1991. Global Geomorphology. Longman,
Harlow.
Summerfield, M.A., 1996. Tectonics, geology and long-term
landscape development. In: Adams, W.M., Goudie, A.S., Orme,
A.R. (Eds.), The Physical Geography of Africa. Oxford
University Press, Oxford, pp. 1 – 17.
Talbot, M.R., 1980. Environmental responses to climatic change in
the west African Sahel over the past 20,000 years. In: Williams,
M.A.J., Faure, H. (Eds.), The Sahara and the Nile. Balkema,
Rotterdam, pp. 37 – 62.
Thomas, D.S.G., Shaw, P., 1991. The Kalahari Environment.
Cambridge University Press, Cambridge.
Tricart, J., 1965. Rapport de la mission reconnaissance ge´o-
morphologique de la valle´e moyenne du Niger. Me´m. Inst.
Fondam. Afr. Noire 72, 193 pp.
Uenzelmann-Nebel, G., 1998. Neogene sedimentation history of the
Congo fan. Mar. Pet. Geol. 15, 635– 650.
Urvoy, Y., 1942. Les basins du Niger. Me´m. Inst. Fondam. Afr. 4,
139 pp.
Ve´tel, W., Le Gall, B., Johnson, T.C., 2004. Recent tectonics in the
Turkana Rift (North Kenya): an integrated approach from
drainage network, satellite imagery and reflection seismic
analyses. Basin Res. 16, 165 – 181.
Wellington, J., 1955. Southern Africa: a geographical study. Physical
Geography vol. I. Cambridge University Press, Cambridge.
Williams, M.A.J., Adamson, D.A. (Eds.), 1982. A Land Between
Two Niles: Quaternary Geology and Biology of the Central
Sudan. Balkema, Rotterdam.
Williams, M.A.J., Williams, F.M., 1980. Evolution of the Nile
Basin. In: Williams, M.A.J., Faure, H. (Eds.), The Sahara and
the Nile. Balkema, Rotterdam, pp. 201 – 223.
A.S. Goudie / Geomorphology 67 (2005) 437–456456