The Lost City Beneath the Amazon Rainforest: What Archaeologists Have Actually Found

For most of the twentieth century, the textbook answer was simple. The Amazon was a “counterfeit paradise,” a lush place where thin soils and punishing humidity kept human societies small and scattered. Real cities, with plazas, roads and planned neighbourhoods, belonged to the Andes or Mesoamerica.

That answer is now crumbling, and the thing doing the damage is a laser.

Over the past few years, airborne lidar has stripped the digital canopy off the rainforest and shown archaeologists something they’d suspected for decades but couldn’t prove: the Amazon was home to genuine urban landscapes, some of them nearly 2,500 years old. So when people search for “the lost city beneath the Amazon rainforest,” the honest answer isn’t one city. It’s a growing list of them.

Here is what we know, what we don’t, and why it matters.

The legend that started the hunt

The idea of a vanished Amazonian civilization is much older than any laser. In 1541, the Spanish expedition led by Francisco de Orellana travelled down the Amazon, and the friar Gaspar de Carvajal, who chronicled the trip, described long stretches of river lined with towns and dense populations. Later explorers found little of the sort, and historians began to wonder whether Carvajal had exaggerated.

Then came the twentieth-century legends. The most famous is the British explorer Percy Fawcett, who became convinced that a ruined city he called “Z” lay in the Brazilian jungle. He disappeared in 1925 while searching for it. His fate spawned decades of speculation, expeditions and at least one bestselling book.

Fawcett never found his city. But the irony is that he was, in a loose sense, onto something. The Amazon did hold large, organised settlements. They just didn’t look like the stone cities he imagined. They were built from earth.

Why nobody could see them

This is the part that surprises people. Amazonian builders mostly worked with soil, not stone. Platforms, mounds, causeways, canals and ditches were all shaped from the ground itself, and when a society collapses or moves on, the forest swallows earthworks fast. Within a few centuries, a plaza can look like an odd rise in the ground under a thick canopy.

Walking through such a landscape, you can stand on top of a monumental platform and never know it. Archaeologists have described exactly this experience. Stéphen Rostain, the French archaeologist behind the Ecuador discoveries, told Live Science that on foot “you have trees in the way,” and it’s hard to see what is hidden. His full comment is quoted in UNSW’s write-up of the research.

Lidar (light detection and ranging) solves that problem. An aircraft or drone fires pulses of laser light at the ground. Most bounce off leaves, but a small fraction slip through gaps in the canopy and reach the soil. Software then filters out the vegetation and builds a bare-earth model of the terrain, revealing every ditch, ridge and platform. It is, in effect, a way of deleting the forest.

The Upano Valley: a “lost valley of cities” in Ecuador

The biggest recent find came from the eastern foothills of the Andes in Ecuador. In January 2024, a team led by Rostain published a study in the journal Science describing a dense network of settlements in the Upano Valley.

Rostain had been working in the region for roughly three decades, so this wasn’t a sudden discovery in the usual sense. He and his colleagues had excavated mounds at sites such as Sangay and Kilamope for years but struggled to see how it all fit together. When Ecuador’s government commissioned a lidar survey, the picture snapped into focus. According to Archaeology magazine’s summary, the data revealed five additional large settlements and ten smaller ones, all densely packed with residential and ceremonial structures.

Some of the details are striking:

• Scale. The survey picked out more than 6,000 earthen platforms, arranged around plazas and along streets.
• Infrastructure. Wide, straight roads ran for tens of kilometres, connecting the urban centres, with smaller streets threading through individual neighbourhoods. The Art Newspaper noted that the roads imply the cities existed at the same time, rather than succeeding one another.
• Farming. Terraces and drainage systems covered the surrounding land. Crops included maize, manioc and sweet potato.
• Age. The Upano people occupied the valley from about 500 BC to somewhere between AD 300 and 600, which makes these settlements around a thousand years older than other known complex Amazonian societies.

Population is where caution is needed. Early press coverage suggested at least 10,000 people, and NBC News reported the same minimum figure, but the researchers have been careful to say the exact number is unknown. The comparison the authors reach for is the “garden cities” of the Classic Maya, where houses sat amid cultivated land rather than in tight stone-built blocks.

“We’re talking about urbanism,” said Fernando Mejía of the Pontifical Catholic University of Ecuador, as Archaeology magazine reported. It is a short sentence, but it sums up how big a shift this is.

Bolivia’s Casarabe culture: cities in the savanna

The Upano Valley isn’t the only candidate for an Amazonian “lost city.” Two years earlier, a team led by Heiko Prümers of the German Archaeological Institute reported something similar from the Llanos de Mojos, a savanna-and-forest mosaic in northern Bolivia.

Their paper in Nature described the Casarabe culture, which flourished from roughly AD 500 to 1400. Lidar mapped about 200 square kilometres across six separate areas and showed a four-tiered settlement hierarchy. Two sites, Cotoca and Landívar, were remarkably large, at 147 and 315 hectares respectively, with monumental platforms and U-shaped pyramids. Smaller settlements surrounded them, tied together by raised causeways and canals that stretched across the landscape. The German Archaeological Institute’s own summary puts the known Casarabe area at about 4,500 square kilometres.

The authors called it “low-density urbanism,” a term already used for tropical cities in Southeast Asia, Sri Lanka and Central America. The idea is that a city doesn’t have to be a compact walled grid. It can be a sprawling, integrated network of centres, farmland and waterworks that functions as a single society. The team described it as the first documented case of this kind in the tropical lowlands of South America.

So is there one “lost city”?

No, and this is where a lot of online coverage gets sloppy. Headlines love the singular. Some of them treat the discoveries as a single mysterious metropolis. The reality is messier and more interesting.

What archaeologists have found is a patchwork of regional societies, each with its own architecture, farming methods and history. Upano and Casarabe are separated by large distances and many centuries. They don’t look alike, and nothing suggests they were connected. If anything, that’s the point. As Rostain put it in the Associated Press coverage, there has always been an incredible diversity of people and settlements in the Amazon, “not only one way to live.”

There’s also a more cautious reading worth keeping in mind. Lidar shows shapes in the ground, not people. Dating, population estimates and social organisation all depend on excavation, which is slow and expensive in a rainforest. Some early claims will probably be revised as the fieldwork catches up with the imagery.

How we got the story wrong for so long

It’s tempting to blame earlier scholars for missing all this, but the story is more complicated. For decades, the dominant view was shaped by the American archaeologist Betty Meggers, who argued that the Amazon’s poor soils limited how large and complex societies could become. It was a reasonable hypothesis given the evidence available at the time, and it held sway partly because the forest hid so much.

Cracks appeared gradually. Researchers working along the Amazon’s tributaries documented terra preta, the dark, fertile soil that ancient communities created by mixing charcoal, bone and pottery into the earth. Terra preta is still productive centuries later, and its presence points to long-term, intensive settlement. Geoglyphs in the southwestern Amazon, enormous geometric ditches that came to light as forests were cleared for cattle, hinted at large-scale community building. Work in the upper Xingu in Brazil suggested networks of villages linked by roads.

Lidar didn’t create this new understanding so much as confirm it and give it dramatic scale. And the surveys aren’t finished. A study published in Science in 2023 combined lidar samples with predictive modelling and estimated that more than 10,000 pre-Columbian earthworks may still be hidden across the Amazon. In other words, we’ve mapped a sliver.

What happened to these cities?

This is the question people always ask, and the honest answer is that we don’t fully know.

For the Upano, researchers have suggested the settlements were abandoned sometime between AD 300 and 600, but the reasons remain unclear. A volcanic eruption from nearby Sangay is among the factors sometimes raised, though that’s a hypothesis rather than a conclusion. For the Casarabe, the decline came before European contact, and again the cause isn’t settled.

What is better documented is the catastrophe that came later. When Europeans arrived in the sixteenth century, they brought epidemic diseases such as smallpox and measles, which spread ahead of the explorers themselves. Population collapse across the Americas was severe, and many historians argue it explains why later travellers found a river valley that seemed so empty. Carvajal’s towns may well have been real; the people who lived in them were gone within a few generations.

Why this matters beyond archaeology

You could file all this under “interesting history” and move on. But the discoveries have real consequences.

They change how we think about the forest itself. If large populations once farmed, managed and reshaped parts of the Amazon, then the idea of a “pristine wilderness” untouched by people needs rethinking. Much of the rainforest may be a cultural landscape, shaped over millennia by human choices. Researchers studying the distribution of useful tree species, such as Brazil nut and palms, have argued for exactly this kind of long-term human influence.

They show how Indigenous knowledge can build sustainable systems. Terra preta, raised fields and managed forests were ways of feeding many people without wrecking the ecosystem. Modern agronomists are studying them, and the lessons may be useful as the region faces deforestation and climate stress.

They add weight to Indigenous land claims and heritage protection. Archaeological evidence of deep, continuous occupation supports what many Amazonian peoples have said all along, and it raises questions about how new sites can be protected from logging, mining and agriculture.

And they remind us how much remains unknown. Remote sensing has made it possible to survey huge areas quickly, but it has also raised ethical questions. Publishing precise site locations can invite looters. Researchers and local communities are still working out how to balance open science with protection.

Can you visit a lost Amazonian city?

Not in the way the phrase suggests. There’s no ruined citadel to wander through, and most of these sites sit on private land, in protected areas or in remote forest where access is limited and sometimes restricted for good reason.

If you’re curious, the better route is through museums and published research. Ecuador and Bolivia both have regional archaeological collections, and the journal articles linked in this piece are freely readable or well summarised. Responsible ecotourism in the Ecuadorian and Bolivian Amazon can also offer a sense of the landscape, but check that any operator works with local communities and respects site protections.

The takeaway

The “lost city beneath the Amazon rainforest” turns out to be a real thing, just not in the form the legends promised. There’s no single golden metropolis waiting at the end of a map. There are roads, plazas, platforms and farms, built from earth by people who knew exactly how to live in this environment and who left their mark on it.

The most exciting part is that we’re early in the story. Each new lidar flight adds detail, and each excavation tests the first impressions. Some of what’s reported today will be refined, and some will probably be overturned. But the direction is clear. The Amazon was never empty, and archaeologists are only beginning to read what’s written in its soil.

Sources and further reading

• Prümers, H., Jaimes Betancourt, C., Iriarte, J., Robinson, M. & Schaich, M. (2022). Lidar reveals pre-Hispanic low-density urbanism in the Bolivian Amazon. Nature.
• Archaeology magazine: Network of Ancient Amazonian Cities Identified in Ecuador
• NBC News: The Amazon’s ancient complex of “lost cities” flourished for a thousand years
• The Art Newspaper: Oldest Amazonian cities uncovered in major archaeological discovery in Ecuador
• German Archaeological Institute (DAI): Lidar reveals pre-Hispanic low-density urbanism