Endorheic basins
A Socratic walk-through of endorheic basins — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why do some of the world's largest rivers never reach any sea, ending instead in a basin with no outlet?
The Volga is Europe's longest river and it reaches no ocean. The Okavango pours out of the Angolan highlands and disappears into a swamp in the Kalahari. The Tarim runs into sand.
We speak of rivers as though reaching the sea were what they are for. But water has no destination — it only goes downhill. So perhaps the question is backwards: rather than asking why some rivers fail to reach the ocean, ask why so many succeed.
Reasoning it through
REASONING #Water flowing downhill stops at a local minimum, not the lowest point on Earth. Any closed depression in the land — a rift floor, a basin between ranges, a crater — is such a minimum. Rivers running into one have nowhere further to go. On that reasoning closed drainage should be the normal case, and drainage to the sea the exception. It is the other way round. Why?
Because a closed depression does not stay closed if you keep pouring. Water fills it until the level reaches the lowest notch in the rim — the sill — and spills over. And spilling water erodes: the outflow cuts the sill down, which lowers the lake, which concentrates the flow and cuts faster. That runaway ends with the basin drained and permanently plumbed into the network running to the sea. Wet closed basins are therefore temporary; the humid world integrates its drainage and forgets it had holes in it.
So closure survives only where the water never reaches the sill. That needs a way out of the basin other than over the rim — and there is exactly one of any size. Evaporation.
Now the mechanism becomes quantitative, and rather elegant. A terminal lake loses water from its surface at some net rate per square metre. Total loss is that rate multiplied by the lake's area. Inflow, meanwhile, is set by the catchment, not by the lake. So if the lake is too small it gains more than it loses and spreads out; if too large it loses more than it gains and retreats. Area is the variable that adjusts until loss equals supply.
Put toy numbers on it to see the shape. A river delivering 10 cubic kilometres a year into a basin with a net loss of 1 metre a year settles at 10 / 0.001 = 10,000 square kilometres of water. Cut the river to 8 and the equilibrium area drops to 8,000 — a fifth of the lake gone, permanently, for a fifth less water. On a floor as flat as these basin floors usually are, that is a shoreline retreating over the horizon.
Three consequences follow without any further assumption, and each is checkable.
The lake is salty. Evaporation removes water and leaves everything dissolved in it behind, so a basin with no outlet is a one-way filter: salts arrive and never leave. Given time, any terminal lake becomes brine.
The shoreline is unstable. Because area does the adjusting, and the floor is flat, modest changes in inflow move the water's edge enormously. Hence playas — salt flats that are lake in a wet decade and dust in a dry one — and hence old strandlines cut into hillsides well above the present shore.
And closure is a joint condition on land and climate, not on land alone. The same rim that holds in an arid interior would be overtopped in a wet one. That is why endorheic drainage is concentrated in continental interiors, rain shadows and subtropical dry belts — central Asia, the Great Basin, the Altiplano, interior Australia — rather than scattered at random. It is a large share of the land surface, commonly given as something like a fifth, though published values move with how basins are delimited and I would not lean on the figure.
The analogy
THE ANALOGY #Think of a puddle on a warm patio under a slowly dripping tap. It spreads until its wetted area is wide enough that evaporation carries off exactly what the tap delivers, and then it stops — not because anything blocks it, but because supply and loss have met. Turn the tap down and it does not drain away; it settles smaller. Dissolve a pinch of salt in it and the salt stays behind, drip after drip. And if the patio has a step at one edge, a heavy enough drip raises the water to the step and it runs away down it — and now there is no puddle at all, only a wet trail.
a patio puddle is gone by evening and remembers nothing, whereas a terminal lake holds water for years to millennia over a floor of its own accumulated sediment and salt — so it responds to a change in the tap with a long lag, and its brine and strandlines record the whole history the puddle cannot.
Clarifying the model
THE MODEL #What builds the rim is a separate matter from what keeps it closed: tectonic subsidence between ranges, volcanic or landslide damming, alluvial fans and dunes blocking an outlet, glacial derangement of an older network. Aridity does not dig the hole; it only stops the hole from being drained. The tempting shorthand — aridity causes closed basins — gets that backwards, and the strong correlation makes the error easy.
What would show this wrong? If closure were about topography alone, closed basins should appear across all climates in proportion to how rugged the land is. They do not. And the sill argument makes a sharper prediction: basins far wetter in the past should carry abandoned overflow channels notched into their rims, together with high shorelines — Lake Bonneville's outlet at Red Rock Pass is the textbook case, as I recall it. Find a basin with unambiguous evidence of a lake standing far above its present level, no sill channel and no strandlines, and this account is in trouble.
One caution against reading the land as destiny. The physics fixes the equilibrium; it does not fix where the inflow goes. The Aral Sea's collapse was not climate acting on a basin but irrigation withdrawals upstream, and the same equation then delivered a much smaller area with complete indifference. In a closed basin, whoever controls the inflow controls the shoreline.
A picture of it
THE PICTURE #How to readStart at Filling and note its two exits. If the rising water reaches the sill it spills, the outlet erodes, and the basin leaves the diagram for good — that is why humid closed basins do not last. Otherwise the lake stops growing where its area makes evaporation equal inflow, and the remaining states are the arid case oscillating: nothing here dries because water drained away, only because it left through the surface.
What became clearer
WHAT CLEARED #A river ending in a basin is not a river that failed. It is a drainage system whose only exit is upward, and its lake is simply as large as evaporation needs it to be to carry off what the catchment delivers. Whether the basin stays closed is decided jointly by rim and climate, since a wet enough climate fills it to the sill and the overflow then destroys the closure for good. The brine, the retreating shorelines and the salt flats all fall out of the same one-way accounting.
Where to go next
ONWARD #- How the salt left behind builds the evaporite sequences that make some of these basins economically valuable.
- Why the sill overflow of a large pluvial lake can produce one of the largest floods in the geological record.
Key terms
TERMS #| Term | What it means |
|---|---|
| Endorheic | of a drainage basin having no outflow to the ocean, its water leaving only by evaporation and seepage. |
| Sill | the lowest point on a closed basin's rim, where a rising lake would first spill out. |
| Terminal lake | the water body at the end of an endorheic system, whose area adjusts until evaporation balances inflow. |
| Playa | the flat, salt-crusted floor of a terminal basin that holds water only intermittently. |
Every term the collection defines is gathered in the glossary.