The strait that wasn't there

This is the waterway in the headlines. For most of the seven months since the war with Iran began in late February, it has been closed to most shipping. On 25 September Iran's foreign minister said Tehran had offered, through Qatar, to reopen it within seven days if the United States lifted its naval blockade of Iranian ports and met other conditions. On 26 September President Trump said: “I reject their proposal.”

It is the only way in or out of the Gulf by sea. In normal times ships keep to two lanes, each two nautical miles wide, one in and one out. In 2024 about 20 million barrels of oil a day went through, about a fifth of all the oil the world uses.

35 maverage depth of the Gulf

Now look at the colours of the water. The Gulf behind the strait is shallow: 35 metres deep on average. Stand a fourteen-storey building on the bottom and its top floors would be in the air. Even near the entrance the floor is only about 100 metres down.

Past the strait, the Gulf of Oman drops past 2,000 metres. That is the dark blue. Everything strange in this story follows from the difference.

17%of the Gulf floor dry

Lower the sea by ten metres. A sixth of the Gulf floor is already dry land. The pale line is today's coast; everything green-brown inside it was seabed a moment ago. Bahrain stopped being an island at five metres.

50%of the Gulf floor dry

At 31 metres, half the Gulf is gone. Qatar has become a hill on a plain joined to Arabia, and Bahrain is part of the same plain. What sea is left is a wide arm along the Iranian shore.

80%of the Gulf floor dry

At 60 metres, what is left is a thin arm of sea along the Iranian coast, following the deepest groove in the floor.

−84 mthe sea leaves the Gulf

Once the sea drops below 83 metres, the ocean pulls back out of the Gulf, because the floor of the narrows is higher than that. Only the deepest hollows inside are still lower than the sea; the pale patches mark them. The ocean now stops at the mouth. For the first time in this descent, there is no strait: the Gulf of Oman ends in a beach.

• hollows lower than the sea

20,000years ago

The last ice age. So much water is locked up in ice sheets that the oceans stand 121 metres below today. The Gulf does not exist. In its place: 237,000 square kilometres of dry land.

• land that was seabed
• today's coast
• rivers the ground implies

Let rain fall on this land and run downhill, and the ground draws its own rivers. The blue lines are that: water following the shape of today's seabed.

They gather into one trunk that starts where the Tigris, the Euphrates and the Karun reach the sea today, runs about 1,200 kilometres down the middle of the basin, and leaves through the Strait of Hormuz. The archaeologist Jeffrey Rose calls it the Ur-Schatt, the ancestor of today's Shatt al-Arab. In thirteen cores taken from the Gulf floor, the lowest layer is river silt laid down between 30,000 and 12,000 years ago.

How big is it? Here is Great Britain at the same scale, turned to lie along the valley. The dry floor, at 237,000 square kilometres, is bigger. Rose's own estimate of the land exposed, about 200,000 square kilometres, puts it at roughly the size of Great Britain.

• Great Britain, same scale

Put today's cities on it. Kuwait City is about 1,000 kilometres from the open sea as the crow flies. Doha is 500 kilometres away, Dubai 120. You could walk from Iran to Qatar with rivers to cross, but no sea.

At the strait itself, the river runs out through a valley where the tankers now sail. Just past the narrows the floor drops into a hollow about 215 metres deep, lower than even the ice-age sea. Lambeck's model lets the late ice-age sea into this hollow; with the river running through it, it could also have been a lake. We found no dated core from it to tell.

• low basin

People lived around this basin long before the ice age peaked. The red dot is Jebel Faya, near today's Sharjah, where stone tools about 125,000 years old were found.

Rose argues that in the driest times the valley, with its river and springs, was a refuge he calls the Gulf Oasis. It is a hypothesis, and not everyone reads the ice age that way: Lambeck writes that it appears to have been cooler and drier than today, with perhaps little chance for lakes to form in the hollows. The floor itself has never been dug: it is under the sea.

• Jebel Faya, stone tools about 125,000 years old

Water may also have come from below. Fresh groundwater still seeps up through the Gulf floor today; a spring off Jubail was still being used in the 1950s. A 2026 study, not yet peer-reviewed, models the areas of the exposed floor where springs could have flowed at the ice-age low: about 13,000 square kilometres on the Arabian side, a figure the authors say could be off by as much as a factor of ten.

Then the ice melts. Around 14,500 years ago the sea jumps: in a few centuries it rises by about 20 metres, a surge known as Meltwater Pulse 1A.

14,000years ago: the strait is back

Around 14,000 years ago, in the middle of that surge as the Barbados corals record it, the rising sea tops the floor of the narrows and pours into the valley. The Strait of Hormuz exists again.

That date is this page's own reading: the moment the coral record of sea level passes the depth of the narrows in today's seabed. It is probably on the late side. The corals grew in the open ocean; in the Gulf the sea stood somewhat higher, and the geophysicist Kurt Lambeck's model of the Gulf opens a narrow waterway at Hormuz earlier, around 17,000 years ago once his radiocarbon dates are turned into calendar years.

By 9,400 years ago half the floor is under water again. On flat ground a small rise moves the coast a long way: Rose cites models in which the shoreline advanced a kilometre a year or more in places. Three dated cores show the sea arriving in the western Gulf from about 11,500 years ago and the water turning fully marine around 8,800.

Lambeck's model has the sea reaching roughly today's coast around 7,000 years ago; the coral record, which this map follows, is still about 7 metres lower then. Between 7,500 and 6,500 years ago, the number of known sites on the Gulf's shores jumps from about ten to more than sixty. The gold dots are some of them. At H3 in Kuwait, from 7,500 to 7,000 years ago, pieces of bitumen have been read as the waterproof coating of boats.

Were these the valley's people, pushed uphill by the water? That is Rose's hypothesis, and the spike in sites fits it. A hypothesis is not an excavation, though, and nobody has dug the drowned floor.

• coastal sites, 7,500–6,000 years ago

Here is the whole return on one chart, from the record this page uses.

Sea level from dated corals: Barbados (Peltier and Fairbanks 2006) up to 8,900 years ago, Tahiti (Bard and others 1996) after that; the dots are the fitted points. These are levels in the open ocean far from the Gulf; in the Gulf itself the ice-age sea stood somewhat higher, because the land there rose and sank differently.

Put the dates in a row. People were using the cave at Lascaux, in France, about 21,000 years ago: that is when the charcoal, bone and antler they left behind date to. At that time there was no Strait of Hormuz. It came back around 14,000 years ago, and the Gulf reached roughly its present shape around 7,000.

The waterway the world is arguing about this week is younger than the Lascaux occupation, and the Gulf in roughly its present shape is not even twice as old as the pyramids.

And here are the tankers again, in their lanes. Under them, through the narrows, runs the valley the river cut on its way to the ocean.

A river valley for tens of thousands of years. A strait for about 14,000. The next time it fills your screen, look at the water and see both.