A Hidden World Beneath the Earth’s Surface

In 1963, a man in the Turkish town of Derinkuyu was renovating his home when he noticed something odd: his chickens kept disappearing into a crack in the basement wall. When he broke through the wall, he did not find a nest. He found a tunnel. That tunnel led to another. And another. Eventually, he had stumbled into an underground city that extended roughly 85 metres below the surface, with enough space to shelter thousands of people along with their livestock and food stores.

Derinkuyu is not unique. Across the ancient world, from the volcanic tuff of Cappadocia to the limestone plateaus of northern France, communities carved entire settlements into the earth. They built kitchens, chapels, stables, granaries and defence systems deep underground. Some of these places could hold thousands of people for weeks or months at a time.

The obvious question is why. Why would anyone invest enormous labour into digging a city underground when the surface was available? The answers are not simple, and they are not the same everywhere. But they all point to a basic truth about human survival: when the surface becomes dangerous, the earth can become a refuge.

Why Ancient People Chose to Live Underground

The reasons vary by location, but they cluster around three pressures: defence, climate and geology.

Defence is the most commonly cited explanation. Cappadocia sat at the crossroads of competing empires. Raids, invasions and religious persecution were recurring threats. An underground city that could be sealed from the inside, with narrow passages that forced attackers to enter one at a time, offered a level of protection that a surface village could not match.

Climate is the second factor. Underground spaces maintain a relatively stable temperature year-round. In Coober Pedy, where surface temperatures can exceed 50°C, underground homes stay around 23°C without air conditioning. In Cappadocia, the same principle applied in reverse: underground rooms stayed warm in winter and cool in summer.

Geology made it possible. Cappadocia’s soft volcanic tuff is easy to carve but hardens on exposure to air. Northern France’s chalk limestone is similarly workable. Without the right rock, these cities could not have been built with the tools available at the time.

A 2026 paper in Zenodo argues that many of these complexes were not designed for permanent habitation at all. They were “temporary buffers” built for short-duration, high-intensity disruptions—weeks to months, not generations. That framing helps explain why the cities were so heavily engineered for defence and storage, yet often abandoned once the threat passed.

Derinkuyu: An Entire World Beneath Cappadocia

Derinkuyu is the deepest excavated underground city in Turkey. It extends to approximately 85 metres and could accommodate up to 20,000 people along with their livestock and food stores. It had wine and oil presses, stables, cellars, storage rooms, refectories and chapels. Starting between the third and fourth levels, a series of staircases led to a cruciform church on the lowest level.

The engineering is what makes Derinkuyu remarkable. More than 50 ventilation shafts supplied fresh air throughout the complex. The largest of these, a 55-metre shaft, appears to have doubled as a water well, providing water to both the surface village and the underground population during a siege.

Defence was built into the architecture. The city could be closed from the inside with large rolling stone doors, and each floor could be sealed off separately. The passages were deliberately narrow, forcing any intruder to enter in single file and at a crouch.

Kaymakli and Özkonak: Different Approaches to Underground Survival

Kaymakli, about 20 kilometres from Derinkuyu, is often described as a smaller sibling. But it has its own character. Its ancient name was Enegup, and its caves may have first been built by the Phrygians in the 8th–7th centuries BC. The city is organised around approximately 100 tunnels, many of which are still used by villagers today as cellars, storerooms and stables.

Kaymakli’s passages are notably low, narrow and sloping. Its kitchens were shared rather than private, and living rooms had underfloor hollows for storing grain. The third floor contains the most important areas: storage places, wine and oil presses, and kitchens.

Özkonak, discovered much later—in 1972, by a farmer named Latif Acar—offers a different design philosophy. Instead of the deep vertical shafts of Derinkuyu, Özkonak uses narrow, elongated openings between floors that serve both communication and ventilation. When the rooms were sealed, air still circulated through holes as narrow as 5 centimetres. The city also has a pipe communication system reaching each of its ten levels, a feature unique among the Cappadocian underground cities.

Özkonak’s defensive architecture is equally distinctive. Holes above the tunnels allowed defenders to pour hot oil or sand onto attackers below. The city could house up to 60,000 people for three months, according to local estimates, though only four of its ten floors are open today.

Petra and Naours: Other Ways People Used the Earth

Not every “underground city” is actually underground. Petra, the rock-cut capital of the Nabataeans in southern Jordan, is often grouped with subterranean settlements, but that label is misleading. Petra is a city carved into cliffs and rock faces, not a settlement built beneath the ground. UNESCO describes it as “half built, half carved into the rock”. Its famous facades—the Treasury, the Monastery, the royal tombs—are monumental rock-cut architecture, not tunnels.

What Petra does share with the underground cities is engineering ingenuity under environmental pressure. The Nabataeans built an elaborate water management system of channels, tunnels, diversion dams, basins and cisterns to capture and store seasonal rainfall in an arid region. That hydraulic engineering enabled a large settled population in a desert environment, much as ventilation and storage enabled underground life in Cappadocia.

Naours, in northern France, is a different case again. It began as a quarry in antiquity and was gradually enlarged over centuries into a refuge network of 28 galleries and 300 rooms, capable of sheltering around 3,000 people at its height. Its purpose was primarily defensive: local populations used it to hide from raiders and armies during the Hundred Years’ War and the Thirty Years’ War. The chimneys from its underground ovens were routed through existing surface structures to conceal the presence of life below.

How Did Thousands of People Survive Underground?

The logistics of underground life were formidable. Air, water, food and waste management all had to be solved.

Air. Ventilation was the most critical constraint. Derinkuyu’s dozens of shafts created a passive airflow system that replenished oxygen across multiple levels. Özkonak’s narrow communication holes maintained circulation even when the city was sealed.

Water. The 55-metre shaft at Derinkuyu served as both a ventilation channel and a well, providing water to the surface village and the underground population. Other cities had their own wells and cisterns.

Food. Storage was built into the architecture. Kaymakli’s living rooms had underfloor grain pits. Derinkuyu had cellars, wine presses and oil presses. Livestock were kept underground in stables.

Waste and duration. The 2026 Zenodo analysis argues that ventilation capacity, moisture accumulation and sanitation challenges imposed a finite operational window—weeks to months, not years. These were refuges, not permanent homes for most inhabitants.

The Engineering Behind Ancient Underground Cities

The builders of these cities were not primitive cave dwellers. They were skilled engineers working with a sophisticated understanding of geology, airflow and structural stability.

The choice of rock was deliberate. Cappadocia’s tuff is soft enough to carve with simple tools but hardens on exposure to air, creating durable walls. The rolling stone doors—some weighing up to half a ton—were engineered to seal passages quickly and could be operated by a single person from the inside.

Water management at Petra was equally advanced. The Nabataeans cut channels into solid rock, built reservoirs and cisterns, and designed diversion dams to control flash floods. Recent archaeological work continues to reveal the sophistication of their hydraulic engineering, revising earlier assumptions about what they understood.

At Naours, the routing of chimney smoke through surface structures shows an understanding of concealment that goes beyond basic excavation. The city was not just dug; it was hidden.

Myths vs Archaeological Evidence

Popular accounts of underground cities often exaggerate their scale, purpose and antiquity. It is worth separating what the evidence supports from what has been embellished.

The claim that Derinkuyu could hold 20,000 people is widely repeated, including by Wikipedia and Turkish cultural sources. The figure is plausible but should be treated as an estimate, not a census. The deeper claim that these cities were permanently inhabited for generations is less well supported. The architectural evidence—heavy emphasis on defence, storage and sealing—points to periodic use during crises rather than continuous occupation.

The dating is also uncertain. Derinkuyu is often attributed to the Phrygians in the 8th–7th century BC, but the evidence for this is indirect. Kaymakli may have Hittite-era origins, over 1,000 years before Christ, but the dating is debated. Özkonak’s builder is unknown, though it is probably Byzantine.

The Blue Zones of underground cities, in other words, do not exist. Each site has its own history, its own builders and its own reasons. Treating them as a single phenomenon obscures more than it reveals.

Why Underground Living Still Exists Today

The most striking modern parallel is Coober Pedy, a town in the South Australian outback where about half of the roughly 2,000 residents live underground. The reason is straightforward: surface temperatures can reach 52°C, while underground homes—called dugouts—stay at a near-constant 23°C. No air conditioning is required. Ventilation shafts run to the surface to supply oxygen and release moisture.

The differences from ancient underground cities are as important as the similarities. Coober Pedy’s dugouts are modern homes with electricity, plumbing and internet. They were built for comfort and energy efficiency, not defence. The residents are not hiding from invaders; they are avoiding the heat.

Naours also had a second life. During the First World War, its galleries became a popular sightseeing destination for soldiers on leave, and they left behind the highest concentration of graffiti from that conflict yet discovered. During the Second World War, occupying Nazi forces used it as a headquarters. A refuge built for medieval warfare became a tourist site, then a military installation, then a museum.

What These Ancient Cities Teach Us

The underground cities of the ancient world are not just archaeological curiosities. They are case studies in adaptation under pressure.

They show that architecture is never only about shelter. It is about threat models, resource management and the social organisation needed to survive. Derinkuyu’s rolling stone doors and narrow passages encode a specific fear: invasion. Coober Pedy’s dugouts encode a different one: heat. Naours’ hidden chimneys encode a third: detection.

They also show the limits of engineering. The same features that made these cities defensible—sealed doors, narrow passages, limited ventilation—also made them difficult to inhabit for long periods. The Zenodo analysis calls this the “Derinkuyu Paradox”: why invest so much labour in a place unsuitable for permanent habitation? The answer may be that permanent habitation was never the point. The point was survival until the threat passed.

Modern architecture is rediscovering some of these principles. Passive cooling, earth-sheltered buildings and underground spaces are being revisited as energy costs rise and climate extremes intensify. The ancients did not have electricity, but they understood something about thermal mass and airflow that modern engineers are still learning to apply at scale.

The underground cities were not escapes from the world. They were ways of staying in it when the surface became too dangerous to occupy.

Conclusion: The Earth as a Refuge

Derinkuyu, Kaymakli, Özkonak, Petra, Naours and Coober Pedy are not the same kind of place. Some are carved from volcanic tuff, others from limestone or sandstone. Some were built for defence, others for climate, others for trade. Some were permanently inhabited; many were not.

What they share is a willingness to go underground when the surface stopped working. That willingness required engineering, social organisation and a clear-eyed understanding of what the environment could throw at a community.

The ancient underground cities are a reminder that human settlement has always been shaped by risk. We build where we can, how we must, and sometimes—when the stakes are high enough—we build downward. The earth, in the end, is not just something we stand on. It is something we can retreat into.