Drowned Worlds — Atlas Essay — 2026-08-10
Before the Persian Gulf
Arabia · Iran · Mesopotamia
Twenty-one thousand years ago, the Persian Gulf was a river basin between Arabia and Iran. The ocean entered through Hormuz and travelled more than a thousand kilometres toward Mesopotamia.
Rising seas turned the exposed river plain between Arabia and Iran into the Persian Gulf.
In brief
- Rising seas turned the exposed river plain between Arabia and Iran into the Persian Gulf.
- At 21,000 BP, ocean level lay roughly 130 metres lower and the floor of the modern Gulf stood above the sea; rivers drained southeast through the exposed basin.
- The Strait of Hormuz opened as a narrow waterway around 14,000 BP, and marine water began entering the central basin around 12,500 BP.
- The Gulf formed over thousands of years, quickening during rapid sea-level rise and each time water crossed a new threshold.
- By about 6,000 BP the shoreline had approached its modern position; near 5,000 BP relative sea level locally rose above the present mark before deltas and coastal flats built outward again.
A gulf-shaped plain
The Persian Gulf is shallow enough to conceal its own recent beginning. Its average depth is only about thirty-five metres. Much of its western and southern floor lies shallower still, crossed by banks, hollows and old channels. Only toward the Strait of Hormuz does the basin deepen into the trough that joins the Gulf of Oman. Lower the ocean by the amount locked in the great ice sheets at the Last Glacial Maximum and almost the entire modern Gulf becomes dry land.
The resulting country was not a level salt pan. Arabia and Iran formed its high walls, but the floor contained ridges, depressions and river valleys. The Tigris, Euphrates and Karun gathered at the northwestern end. Wadi systems descended from Arabia. Along the Iranian side, short steep rivers carried water and sediment from the Zagros. Their combined drainage followed the basin southeast toward Hormuz. Marine geologists sometimes call the master stream the Ur-Schatt: not a lost city-bearing river recovered in a trench, but a useful name for the ancestral drainage that occupied the axis now buried beneath marine sediment.
This geometry explains the scale of the transformation. On a steep coast, a vertical rise of ten metres may move the waterline only a short distance. Across the Gulf floor, the same rise can cross broad low country. A sea advancing from Hormuz therefore did not merely trim the margins of Arabia and Iran. It replaced the landscape between them.
The rivers beneath the sea
Ice Age Arabia was often dry, but dry does not mean empty of water. The exposed basin collected the discharge of a catchment extending far beyond its shore. The Euphrates and Tigris carried snowmelt and rain from Anatolia and the Zagros. The Karun brought water from southwestern Iran. Aquifers beneath the Arabian Peninsula issued through springs at lower elevations. During wetter intervals, lakes and marshes occupied depressions on the basin floor; during harsher intervals, the great rivers and spring lines remained the dependable structure of the landscape.
Jeffrey Rose calls this well-watered country the “Gulf Oasis.” When aridity squeezed people out of much of Arabia, the exposed basin offered fresh water and a lower, cooler corridor between southern Iran and the peninsula. As the sea removed that corridor, settlements multiplied around its retreating edges.
No village has yet been excavated from the Gulf floor, whose old surfaces now lie beneath water and thick marine sediment. The plain nevertheless lay between occupied regions, gathered the fresh water of an immense catchment and offered the rivers and springs by which people crossed arid country.
The ocean enters at Hormuz
The filling began at a gate. As ice melted, the ocean rose while the crust responded to the loss of ice and the growing load of water. The Strait of Hormuz opened about 14,000 BP as a narrow passage into the lowest southeastern reaches. Around 12,500 BP, marine water entered the central basin. From there the coastline moved northwest.
Global ice volume is near its maximum. Nearly all of the modern Persian Gulf floor is exposed as a river basin.
Deglacial rise begins to draw the Indian Ocean toward Hormuz while the Gulf basin remains predominantly terrestrial.
Meltwater Pulse 1A accelerates global sea-level rise and the Gulf transgression.
Water opens a narrow Strait of Hormuz and enters the southeastern depression.
The marine incursion reaches the central basin; the new inlet lengthens rapidly along the old drainage axis.
Successive low shelves and river valleys drown as the shoreline continues northwest.
The sea occupies most of the modern Gulf basin, while broad coastal plains remain around its head and margins.
The shoreline approaches its modern position at the northwestern head of the Gulf.
Relative sea level reaches a local mid-Holocene highstand in parts of the Gulf; deltas and sabkhas subsequently remodel the coast.
Meltwater Pulse 1A raised the world ocean unusually quickly around 14,600 BP, and other intervals of rapid ice loss followed. Each sill crossing opened another low reach to the sea. Even during the fastest intervals, the Gulf filled in stages over thousands of years rather than in one basin-wide flood.
For people living beside it, a millennial process was still a succession of local emergencies. A river mouth became an estuary. A spring turned brackish. A route across a low divide became a channel. One generation inherited a coast farther inland than the coast known by its grandparents. The absence of one basin-wide catastrophe does not make the transformation gentle at the scale of a household.
People followed the moving edge
The oldest archaeological remains around the Gulf do not form a continuous census, but the early Holocene pattern is striking. By the seventh and sixth millennia BCE, communities were living along the coasts of eastern Arabia, Kuwait, Bahrain, Qatar and the United Arab Emirates. They fished, collected shellfish, herded animals, made shell ornaments and crossed open water. At sites such as Dosariyah in eastern Saudi Arabia, occupation coincided with the mid-Holocene transgression. At H3 in Kuwait, bitumen fragments bearing impressions of reeds preserve evidence for watercraft in the sixth millennium BCE. Ubaid-related pottery moved far beyond southern Mesopotamia, showing a Gulf that had become a corridor of boats and exchange.
This was not a simple exodus in which one population marched ahead of the tide. The coast itself offered new resources. Shallow lagoons, mangroves, tidal flats and rich fisheries grew where dry valleys had been. Some communities moved; some changed their seasonal rounds; some may have remained near familiar rivers as the rivers became estuaries. The new sea destroyed one geography and created another.
Water and sediment buried camps on the last Ice Age plain. Later coastal sites may now stand inland because sea level fell slightly, deltas grew, tidal flats accumulated and wind-blown sand altered the margins. Communities repeatedly found new shores as the old ones moved.
The head of the Gulf kept moving
At the northwestern end, sea-level rise met one of the world’s great sediment machines. The Tigris, Euphrates and Karun did not stop at a fixed shore. They shifted channels, built levees, filled marshes and pushed their deltas into the new sea. Tectonics and subsidence changed the relative height of land and water. As a result, the Holocene shoreline of lower Mesopotamia cannot be drawn as one clean arc for every century.
Lower Khuzestan became a shifting mosaic of marine and estuarine channels, wetlands, flood basins and playas as the sea advanced and rivers supplied sediment. Ocean water crossed the whole Gulf basin and reached its northwestern head while deltas built back toward it.
Near 5,000 BP, relative sea level around parts of the Gulf stood about a metre or more above the modern mark. The present coastline followed later relative sea-level fall, sediment accumulation, channel migration and human alteration. The water made the Gulf; rivers immediately began remaking its head.
A sea made in human time
Today Iran and Arabia face each other across an international waterway. At the Last Glacial Maximum, those shores belonged to one drainage basin. Hormuz was a threshold beyond its southeastern end. Rivers ran along country now crossed by tankers.
Over sixteen thousand years, the ocean converted that drainage into a sea. Melting ice raised the water, faster pulses quickened its advance, and the crossing of Hormuz opened successive depressions across the basin.
The human story is equally direct. People occupied the surrounding lands before, during and after the transgression. The exposed basin concentrated the freshwater of an enormous catchment; the new coast then concentrated fish, shellfish, boats and exchange. What disappeared was not blank continental shelf. It was a habitable middle between two regions that the sea later taught us to see as opposite shores.
Rising seas turned the exposed river plain between Arabia and Iran into the Persian Gulf.
Key sources
Kurt Lambeck, “Shoreline reconstructions for the Persian Gulf since the last glacial maximum” (1996), Earth and Planetary Science Letters 142, pp. 43–57, for the regional relative-sea-level history, the opening of Hormuz around 14,000 BP, marine incursion into the central basin around 12,500 BP and the approach to the modern coast before 6,000 BP.
Kurt Lambeck et al., “Sea level and global ice volumes from the Last Glacial Maximum to the Holocene” (2014), pp. 15296–15303, especially fig. 4, for the global ice-volume history and the timing of rapid deglacial rise.
Jeffrey I. Rose, “New Light on Human Prehistory in the Arabo-Persian Gulf Oasis” (2010), Current Anthropology 51, pp. 849–883, especially pp. 849–861, for the exposed basin’s river-and-spring geography and its proposed role as a human refugium.
Gary A. Cooke, “Reconstruction of the Holocene coastline of Mesopotamia” (1987), Geoarchaeology 2, pp. 15–28, for the transgression at the head of the Gulf and the later southward growth of the Shatt al-Arab delta.
Frieda Bogemans et al., “New data on the sedimentary processes and timing of the initial inundation of Lower Khuzestan” (2017), The Holocene 27, pp. 613–620, for facies and dated cores at the northeastern head of the Gulf.
Stephen W. Lokier et al., “Late Quaternary sea-level changes of the Persian Gulf” (2015), Quaternary Research 84, pp. 69–81, for the transgressive phase, mid-Holocene highstand and subsequent relative sea-level history.
Jacques Connan et al., “A comparative geochemical study of bituminous boat remains from H3, As-Sabiyah (Kuwait), and RJ-2, Ra’s al-Jinz (Oman)” (2005), Arabian Archaeology and Epigraphy 16, pp. 21–66, for the impressed bitumen evidence associated with reed boats.