Looking back into Botswana's geological history gives us insights to explain how the Great Salt Pans and the Okavango Delta were formed. It's a long story though, of which I can give only a brief outline here. For a deeper and much more scholarly approach, seek out John Reader's Africa: A Biography of a Continent and especially Mike Main's Kalahari: Life's Variety in Dene and Delta. Both are excellent, giving much more comprehensive discussions of the latest scientific thinking than I'm able to give here – the first in terms of Africa generally, and the second with immense detail specifically about Botswana.

Setting the scene

To put discussions in perspective, and give our history a sense of scale, let me start at the beginning… Around 4,600 million years ago the Earth's crust began to form and cool. A thousand-million years later, fossils from South Africa's Barberton Mountain Land give us evidence of the first recorded simple, single-celled bacteria.

By 3,300 million years ago more complex blue-green algae had appeared, though it took a further 2,300 million years for multi-cellular organisms to evolve. Palaeontologists say that around 600 million years ago there was an 'evolutionary explosion', when many different species evolved, including ancestors of most of the world's existing invertebrates.

About 230 million years ago the Earth saw the emergence of both dinosaurs and the first mammals. For the next 165 million years the dinosaurs dominated the earth and the largest mammals which evolved were probably no bigger than present-day rabbits. Only around 65 million years ago, when some great calamity killed off most of the dinosaurs, did the mammals begin to rise to dominance and inherit the earth.

The Super-continent

Meanwhile somewhere between 135 and 10 million years ago the super-continent, Gondwanaland, broke up and the continents started to move away from each other.

By the time that the continents split from Gondwanaland, the basic rocks upon which southern Africa is built, the Karoo lavas, had been laid down. By about 65 million years ago – the end of the Cretaceous Period – most of the subcontinent's diamonds, gems and other mineral wealth had also been formed, and erosion was gradually wearing away at the land.

The world's earliest primate fossils are from Europe and North America, dating from around 65 million years ago, but it wasn't until about 4 million years ago that Australopithecus appeared in Africa – which is a prime candidate for being one of our earliest ancestors.

This period, as our first ancestors were evolving, is about the time when we start looking at how Botswana's landforms were created.

Kimberlite pipes

Diamond is a crystalline form of ordinary carbon formed under conditions of extreme pressure and temperature. In nature, such conditions are only found deep below the Earth's surface, in the lower crust or upper mantle. Under specific circumstances, the rock matrix in which diamonds occur was subjected to such great pressure that it became fluid and punched its way up to the earth's surface in a volcanic pipe of liquid rock. This situation is similar to a conventional volcanic eruption, except that instead of basaltic magma being erupted through fissures in the crust, the volcanic material is a peculiar rock called kimberlite. This contains a wide assortment of minerals (including diamonds), often in addition to large chunks of other rocks that have been caught up in the whole process.

Such kimberlite pipes occur throughout southern Africa from the Cape to Zaire. However, only a small proportion of those discovered contain enough diamonds to be profitably worked.

Botswana mines kimberlite pipes for diamonds at Orapa and at Jwaneng. It's interesting to note that Botswana's neighbour, Namibia, has diamond deposits along its Atlantic coast. These are 'secondary diamond deposits', because they do not come directly from kimberlite pipes. Instead, Namibia's diamonds are found in areas where ancient rivers have eroded kimberlite pipes in the interior, washing diamonds down to the sea, and depositing them in sediments there.

The Kalahari's superlake

During the Tertiary Period, which dates from about 65 million years ago, Botswana's climate was probably very arid, and as the region's rocks were gradually eroded, so the wind-blow sands accumulated and began to form what we now call the Kalahari.

Around four or five million years ago it's thought that the Okavango, Kwando and the Zambezi rivers had completely different courses than they do today – probably all flowing into one channel which headed south through the Kalahari and into the Orange River and/or the Limpopo. (Different theories suggest different courses for these.)
Then, around 2–4 million years ago, it's thought that seismic shifts forced parts of southern and central Botswana upwards. Geologists identify the areas raised as the 'Zimbabwe-Kalahari Axis' and the 'Bakalahari Schwelle' – both of which are now watersheds in the region.

The net effect of these upward movements was to block the flow southwards of the rivers, and to form a super-lake that had its deepest parts where the Makgadikgadi Pans are today. This greater Lake Makgadikgadi existed until very recently, though its extent varied greatly over the millennia, depending on the climate and inflows. It is thought that at its greatest, it covered an area of up to 80,000km2, and that it probably only dried up in the last 10,000 years. At its largest it probably stretched as far as Lake Ngami in the west, what is now Chobe in the north, and beyond Nata to the edge of present-day Zimbabwe in the east.

Evidence suggesting the existence of this lake is dotted around its ancient shorelines. The great Magwikwe Sand Ridge that you cross as you drive between Savuti and North Gate, probably defined one of its northwestern shorelines. Similarly, another is thought to have been the less obvious Gidikwe Sand Ridge, which lies just to the west of the western border of the present Makgadikgadi National Park.

Look at the base of several of the Kalahari's isolated hills and you'll find rounded, water-washed pebbles and rocks worn smooth over the centuries by waves. The eastern side of the Gchoha Hills, north of Savuti, are a particularly clear example of this – though all of Savuti's hills show some of this history if you look carefully.

In some areas of the Makgadikgadi you can stroll around and pick up the flint arrowheads and other tools of Stone-Age encampments, which must have sat beside the shores of this lake within the last few thousand years.

From rifts to deltas

At some time in the last million years (none of my sources seems to be sure exactly when), this flow into the great Lake Makgadikgadi was gradually, probably quite slowly, cut off.

Northern Botswana has a series of deep, underlying fault lines running beneath its sands. These faults are thought to be the southernmost extensions of the same system of parallel fault lines that are pulling away from each other and have formed East Africa's Great Rift Valley. These extend south, forming the rift valleys of the Luangwa and Lower Zambezi, before cutting across the Zambezi at Victoria Falls. Here amidst the sands of the northern Kalahari they are probably at their youngest and shallowest. In Botswana and Zimbabwe these two main fault lines run parallel to each other, from northeast to southwest.

Look at a map and you'll see the Zambezi flowing roughly south through western Zambia, until it reaches a point around Victoria Falls and what's known as the Middle Zambezi Valley (around where Lake Kariba is today). Effectively the Zambezi has been diverted, flowing into the rift valley and then northeast along it. This would have starved the great lake of its largest water supply, but clearer evidence of the fault lines can be seen in the present paths of two other rivers.

Look again at the maps and you'll realise that the Kwando forms the Botswana-Namibia border flowing southeast (roughly parallel to both the Zambezi and the Okavango). Then it clearly passes through an area where there's very little gradient, and forms a mini-delta, before abruptly changing direction to become the Linyanti River, and starts to flow northeast. (Note, in passing, the very clear parallels between the delta formations of the Okavango and Kwando/Linyanti rivers.)

This new course of the Linyanti River marks one fault line – known as the Linyanti-Gumare Fault. At Lake Liambezi, this river then seems to 'escape' from the confines of its fault to meander roughly southeast, until meeting another fault line parallel to the first. Here it is again renamed as the Chobe River, with a clearly defined course, flowing northeast and parallel to the Linyanti, along another fault line.

The more northerly of these faults is the important Linyanti-Gumare Fault. The Selinda (or Magwegqana) Spillway also flows roughly along this, and further west the line cuts across the flow of the Okavango, just south of the base of the Panhandle. It effectively marks the start of the Delta proper. Parallel to this, but at the south end of the Delta, are the Kunyere and Thamalakane faults. The latter is clearly marked by the position of the Thamalakane River, which marks the most southerly extent of the Delta. Here the Thamalakane River collects the meagre outflow from the Okavango's Delta and diverts it southwest along the fault line, and ultimately into Lake Ngami and the Boteti. Maun sits in the narrow space between the Kunyere and Thamalakane faults.

Just like the parallel faults of the rift valleys further northeast, the Linyanti-Gumare Fault in the north, and the Kunyere and Thamalakane fault lines in the south are gradually pulling away from each other. Between them, the underlying base rock has dropped relative to the land outside the faults – some geologists say by as much as 300m. These have acted to 'capture' the Okavango River in a depression where the gradient is very shallow. (As an aside, just north of this, smaller faults, perpendicular to the main ones, act to restrict the Okavango's sweeping meander to the narrow confines of the Panhandle.)

Thus, in time as they gradually formed, these faults diverted the Zambezi and the Kwando rivers, and forced the Okavango to spread into its present-day delta formation.

The Great Salt Pans

As its feeding rivers were diverted, the great Lake Makgadikgadi was starved of its sources. This would probably have happened over a very long period of time. We know that the climate in Botswana varied greatly, and during periods of heavier rainfalls the lake might have persisted. Geologists have identified at least five clearly different levels of the lake – each of which has its own identifying features that are still visible. And for each the size of the lake must have been totally different.

But eventually, starved of inflow in a drier climate, the lake shrank. With no known outflows, it would probably have already been very brackish. Now its remaining salts were concentrated more and more, and eventually crystallised out where the last of its waters evaporated – at its lowest point, where the pans are now. Too saline for plants to grow, this residue formed the amazingly flat surface of the Great Salts Pans that are now such a distinctive feature of the northern Kalahari.