The Zambezi and the Smoke that Thunders: how rivers change through time


The Zambezi is one of Africa's largest rivers, but that wasn't always so. It used to be a small river on the coast of the Indian Ocean in Mozambique. Like an advancing front the Zambezi captured inland rivers, creating a huge catchment area that reaches almost to the Atlantic Ocean. The natural wonder of Victoria Falls is the living proof of this history. I add a few personal touches to this article, because in another life, when I worked in the petroleum industry, I specialized in the field of palaeodrainage with a focus on Africa. I discovered a few missing links in river evolution ... but in the publishing world, I bumped into scientific misconduct on more than one occasion.

Author : Kathelijne Bonne.

Victoria Falls (K. Bonne)
Victoria Falls (K. Bonne)

The Zambezi River is 2660 km long and rises in the Bié Mountains in Angola, less than 200 km from the Atlantic Ocean. Its upper course in the west, the Upper Zambezi, meanders lazily through the high plains of Angola and Zambia. Then it plunges down in a kilometer-wide water curtain of the Victoria Falls, named after the famous queen of the British Empire by explorer David Livingstone. I prefer the African name though, Mosi-oa-Tunya, or the Smoke that Thunders. I once stood on the great bridge between Zambia and Zimbabwe, soaked by the spray and in awe of thundering smoke. If a normal rainbow is not enthralling enough, at full moon, and when the falls are at their greatest discharge, an other-worldly moonbow is cast in the spray. Beyond the falls, the Mid-Zambezi rages through the Batoka Gorge, a rafters' paradise. To the east, the river runs through the valleys of the East African rift system. There were once rapids, but they have been tamed by two massive hydroelectric dams (Lake Kariba and Lake Cahora Bassa are the associated reservoirs). Downstream, in Mozambique, the Lower Zambezi empties into a delta in the Indian Ocean.

The modern Zambezi came into being because the Lower Zambezi captured the middle and upper courses. This happened over a span of millions of years and was caused by the stepwise breakup of supercontinent Gondwana (watch this good animation of Gondwana), which was accompanied by volcanism and changes in relief. Hence, the rivers had to change too, as we will see.

Zambezi Basin (Eric Gaba/Wikipedia; edits: K. Bonne)
Zambezi Basin (Eric Gaba/Wikipedia; edits: K. Bonne)
Victoria Falls (Diego Delso/Wikipedia).
Victoria Falls (Diego Delso/Wikipedia).

Like a body

Rivers are like veins in a body, redistributing nutrients, organisms, and masses of water with dissolved minerals, salts, sediment, and soil particles. And it is only natural that rivers change their course as time passes. They capture parts of other rivers, or lose part of their watershed. This has implications for the distribution and mixing of fauna and flora - and of genetic material - over time.

Stream capture

Rivers undergo changes as one river captures another and incorporates part of another drainage basin in its own. How does this happen? Firstly, an event that triggers the capture process is needed. This is usually uplift of the land surface. Uplift causes the river to incise its course into the relief, i.e. eroding downward, creating a steeper valley. This in turn causes the valley slopes to recede and the watershed divides to be pushed back, in such a way that the drainage basin becomes larger. At some point, this erosion, also called headward erosion, 'eats into' the basin of the neighboring river, capturing its upper reaches. This phenomenon is common and is called stream piracy or stream beheading.

River expansion

All large river basins have been created by stream piracy, just as large countries have obtained their territory in one way or another by conquest and certainly also by beheading. I take this opportunity to introduce a simpler name for the phenomenon in which river basins become larger through stream beheading or other processes: river expansion. 


To understand past drainage changes, anomalies in the river network and in the relief must be identified. The geology and palaeo-relief, which change through time, must be known as well. Without that background, don't even get started. Palaeodrainage, the science of reconstructing river evolution, is therefore like solving a huge puzzle with many missing pieces.

When I worked for the British exploration company Getech in Leeds, I researched the history of several large rivers, e.g., the Amazon, Indus, Ganges, Niger and Volta (and even the rivers of pre-glacial Greenland!), which all flowed in completely different directions in the past. I unearthed the 'deep' history of the Niger River and its Great Bend, published in 2014 by the Geological Society of London.

Diamonds and petroleum

Why is the exploration industry interested in palaeodrainage? Rivers transport sediment to the oceans, and the type and amount of it helps determine the likelihood of finding hydrocarbons. Diamonds are also redistributed by rivers; due to certain properties, they concentrate in specific fluvial environments. It is hence no surprise that much palaeodrainage research is funded by diamond giants like De Beers. 

When I left Getech, I wanted to take river reconstruction a little further. I chose the history of the Zambezi River because I had already visited that region, and I hope to go there again. I discovered a few missing links in the published but fragmentary Zambezi palaeodrainage. But like any researcher, I watch like a hawk over my little discoveries, and I am not very keen on sharing them. 

Monopoly and piracy

There are many unscrupulous rivals in the field, leaders of research groups who think they have a monopoly over a particular area, basin, river or volcano. These 'experts' often 'contribute' as anonymous reviewers for scientific journals. Conveniently hidden under the cloak of anonymity, they lie in wait to throw other people's manuscripts in the paper bin (read: do everything to prevent publication, and then use the information for their own research). A few names are on the tip of my tongue. Unfortunately, unethical conduct in publishing is all too common. Piracy indeed happens as much on paper as it does in nature or on the seas. Let's return to the Zambezi!

Zambezi River evolution

Abrupt bends, rapids and waterfalls in rivers are usually signs of capture or changes in flow direction. The same is true for the Zambezi. We explain the evolution in three stages, starting at 180 million years ago (the Early Jurassic). That's when the coast of East Africa formed, into which the Zambezi flows. Before that, of course, there were rivers in Gondwana, but they were completely different and that story would take us too far in this article.

Small Zambezi River during the Jurassic.
Small Zambezi River during the Jurassic.

Stage 1: The small Lower Zambezi

180 million years ago, the land surface in Gondwana, where now southeastern Africa lies, began to bulge tremendously. The bulging area was a high plateau at 2000 meters above sea level and it had a diameter of about 2000 km. The uplift was caused by the rising of a 'mantle plume' from deep in the earth (in this case the Karoo Plume). This caused a gigantic amount of magma to flow out over almost all of southern Africa, on a scale that has no contemporary counterpart. Scientists call the resulting volcanic area a Large Igneous Province (LIP). Karoo basalt can be seen today in the Drakensberg Mountains and in Victoria Falls. Because of the uplift, many rivers of southern Africa began to flow from east to west, away from the high area.

In the midst of the volcanic highlands, large rift valleys formed (compare it to a cake in the oven, as it bakes the surface sometimes bulges, only to collapse and crack). Along one such fissure, Antarctica broke away from East Africa. The sea flowed into the valley between the two continents: the Indian Ocean and the coasts of East Africa were born. The coast was very steep (rising to the inland plateau of 2000 meters above sea level). Short but wild rivers raced down. One of those short rivers was the young Zambezi. It happened to be located in a place where the rocks were extremely fractured. Therefore, it eroded with more ease and expanded towards the interior. This is how the Lower Zambezi came into being. But it remained, for the time being, a fairly modest river. 

mega-Limpopo River during the Cretaceous.
mega-Limpopo River during the Cretaceous.

Stage 2: The great Limpopo River

135 million years ago (Early Cretaceous), the west coast of southern Africa was uplifted by the Tristan mantle plume. Once more, a bulge formed in the landscape and lava flowed generously. South America tore away from Africa, creating the Atlantic Ocean. The supercontinent Gondwana was no more. And with the land surface rising in that place, a slope developed from west to east, opposite to the earlier direction inflicted by the Karoo Plume. The rivers also changed direction and a large eastward flowing river network developed. Amongst those great tributaries were the Upper Zambezi and Okavango rivers. These rivers linked with the Limpopo River in the southeast, which was then the largest river in southern Africa (the 'mega-Limpopo'). But that configuration did not last.

A kind of bowl-shaped depression formed in the interior of southern Africa: the Kalahari-Congo Basin. The Upper Zambezi River emptied into the middle of that bowl, creating an ecosystem of wetlands, lakes and sandy plains. This situation persisted for a long time. The Upper Zambezi was still far from connecting with the small Lower Zambezi that flowed to the sea in the east.

Stage 3: One great united Zambezi River

The Zambezi captures many interior rivers.
The Zambezi captures many interior rivers.

Africa was shaken again as the East African Rift extended like a zipper from Ethiopia to Mozambique and Botswana throughout the last 30 million years. In the Zambezi region, linear rift depressions formed where today the Kariba and Caborra Bassa reservoirs are located (the Gwembe, Caborra Bassa and Mana Pools basins). As surface water logically flows downhill and collects in depressions, a great stream developed along the axis of these basins: the Middle Zambezi. It drained to the east, and connected with the Lower Zambezi, forming one greater Zambezi River. This was the first step in the river expansion of the modern Zambezi.

Because the Middle Zambezi in the rift valley was at a lower elevation than the surrounding plateaus, erosion took place at the edges. And through headward erosion, it captured other major watercourses and its catchment area grew larger and larger. Among the 'conquests' are the Shire River (the drainage of Lake Malawi) and the great Luangwa River. And eventually it also captured the Upper Zambezi which until then flowed quietly to the middle of the Kalahari Basin. This last linkage created the Victoria Falls. These waterfalls are actually the active 'front' of the headward erosion, which continues to this day. 

Victoria Falls in the future?

How the rivers flow today is just a snapshot in time. Everything continues to move, almost imperceptibly on a human scale. The Victoria Falls are slowly but surely receding, migrating upstream. Several scenarios are possible in the future. The Zambezi may drain the region of the Okavango Delta, if the falls recede far enough. Then the Delta will cease to exist. But it doesn't have to happen that way. The Victoria Falls can also dry up, then erosion will stop. This could be due to climate change, but also due to the creation of a deep rift valley in the Okavango region. Perhaps then a large lake will form there, like the Malawi and Tanganyika lakes. But now I am getting too far ahead of myself. Whatever will happen, and in whatever direction the rivers will flow, nature will be as magnificent there as it is today.


Images, maps, pictures: @Kathelijne Bonne, unless otherwise stated. 

Gondwana video: Earthworks by Colin Reeves. 

Title video: Tom Varley - Victoria Falls, Zimbabwe: 

A few sources: 

Bonne K., 2016, Streamstar Geosolutions, Zambezi: a new palaeodrainage evolution

Bonne & Caracciolo, Prediction of volumetric data and reservoir properties of the Mozambique Margins, from sediment source to sink reconstruction using an empirical model and analogues , International Meeting of Sedimentology, Toulouse, France. Oct 10, 2017.

Bonne K., 2014 Reconstruction of the evolution of the Niger River and implications for sediment supply to the Equatorial Atlantic margin of Africa during the Cretaceous and the Cenozoic. Geological Society of London Special Publications 386.

Burke, K., & Gunnell, Y. (2008). The African Erosion Surface: A continental-scale synthesis of geomorphology, tectonics and environmental change ove rthe past 180 million years. Boulder, Colorado, USA: The Geological Society of America. Memoir 201, p. 66.

Burke, K., Macgregor, D., & Cameron, N. R. (2003). Africa' s petroleum systems: four tectonic "Aces" in the past 600 million years. In T. J. Arthur, D. Macgregor, & N. R. Cameron (Eds.), Petroleum geology of Africa; New themes and developing technologies, Geological Society, London, Special Publications 207, 21-60.

Cox, K. G. (1989). The role of mantle plumes in the development of continental drainage patterns. Nature, 342, 873 - 877.

De Wit, M. C. J. (1999). Post- Gondwana drainage and the development of diamond placers in western South Africa. Economic Geology, 94, 721-740.

Goudie, A. S. (2005). The drainage of Africa since the Cretaceous. Geomorphology, 67(3-4), 437-456.

Moore, A., & Blenkinsop, T. (2002). The role of mantle plumes in the development of continental-scale drainage patterns: The southern African example revisited. South African Journal of Geology, 105 (4), 353-360.

a few keywords: zambezi river, victoria falls, evolution of the zambezi river, river expansion, zambezi palaeodrainage, stream capture, stream piracy, zambezi history, zambezi river history, victoria falls history, gondwana drainage, smoke that thunders, unethical review.