Origins of Civilisation · Upano Valley, Morona-Santiago, Ecuador

Upano Valley (Sangay & Kilamope)

Six thousand earthen platforms and twenty-five kilometres of dead-straight road, hidden under Amazon canopy for two millennia — the oldest urbanism yet found in the rainforest.

Mainstream: c. 500 BC – AD 300/600 (Kilamope and Upano phases; later Huapula reoccupation)Alternative: Date not disputed — the argument is over what the word 'city' is doing, and over who gets to claim the discovery as vindication-2.13°, -78.11°

Origins Engineering Lost Worlds

At a glance

The Upano Valley runs along the eastern flank of the Ecuadorian Andes, where the mountains break down into Amazonian forest beneath the perpetually erupting Sangay volcano. In January 2024, a team led by the French archaeologist Stéphen Rostain published a LiDAR survey of some 300 square kilometres of that valley in Science, and it revealed a landscape nobody had expected: more than 6,000 rectangular earthen platforms, most about 20 by 10 metres and standing two to three metres high, grouped in clusters of three to six around open plazas. Five large settlements and ten smaller ones were laid out across the valley floor, stitched together by wide, straight roads running as far as twenty-five kilometres and cut through the terrain rather than around it, alongside terraced fields and drainage channels. Radiocarbon dating and long-running excavation put the construction between roughly 500 BC and AD 300–600, the work of the Kilamope and Upano cultures, with a later Huapula reoccupation. That makes Upano the earliest known urbanism in Amazonia by something over a thousand years. The site had not been entirely unknown — settlement traces were noted in the 1970s, and Rostain himself had excavated at Sangay and Kilamope for twenty-five years — but nothing visible on the ground conveyed the scale. Only when the canopy was stripped away in the LiDAR point cloud did the valley resolve into what Rostain called an entirely human-engineered landscape.

The mainstream view

What archaeology says

The interpretation offered by Rostain, Antoine Dorison, Fernando Mejía and their colleagues is 'garden urbanism': a low-density, agrarian form of city in which monumental platform groups and ceremonial plazas are embedded in continuous cultivated ground rather than packed behind walls. Maize, beans, sweet potato and manioc grew in the spaces between complexes. The roads are the load-bearing evidence for reading the network as a single functioning system rather than a scatter of unrelated villages: they are engineered, they run between the centres, and they imply the settlements stood and were used at the same time. 'We're talking about urbanism,' as Mejía put it.

The dating rests on stratified excavation rather than on the LiDAR alone, which is the crucial methodological point. Profiles at Kilamope show multi-phase mound construction with domestic floors, postholes and hearths above basal layers around 500 BC; refuse pits at Sangay have yielded radiocarbon determinations spanning roughly 165 BC to AD 655. Ceramic sequences shift recognisably across the phases. The volcano that made the valley habitable — Sangay's ash keeps the soils exceptionally fertile — is also Rostain's candidate for why the system eventually failed.

Reception among specialists has been enthusiastic rather than sceptical, but not uncritical. Thomas Garrison, a LiDAR specialist at the University of Texas at Austin who was not involved in the work, cautioned that it is too early to set Upano beside the Classic Maya or Teotihuacán, which were considerably more complex and more extensive. The disagreement, where it exists, is about how much weight the word 'city' can carry — not about whether the platforms, the roads and the dates are real.

Key evidence cited

2015 LiDAR survey of c. 300 km2, published by Rostain et al. in Science (January 2024), mapping 6,000+ platforms across 15 settlements
Twenty-five years of excavation at Sangay and Kilamope showing stratified mound-building over domestic floors, postholes and hearths
Radiocarbon dates from Sangay refuse pits spanning c. 165 BC – AD 655, anchoring the network in excavation rather than in the LiDAR alone
Engineered straight roads up to 25 km long linking the centres, implying the settlements were contemporaneous and functionally integrated
A continuous local ceramic sequence across the Kilamope, Upano and Huapula phases, with no intrusive material culture

The alternative view

What the skeptics propose

Upano belongs to the small class of sites where the heterodox instinct was broadly right and the textbooks were wrong — the same story the Acre geoglyphs tell. For most of the twentieth century, Betty Meggers's 'counterfeit paradise' orthodoxy held that Amazonian soils could not sustain large complex societies. Graham Hancock, among others, spent years arguing the opposite: that the Amazon concealed great cities beneath its canopy, that the early Spanish chronicles of dense riverside populations deserved to be believed, and that the basin was one of the vast unexamined regions where the archaeological consensus had simply never looked. On that narrow claim — the forest hid substantial ancient urbanism — the LiDAR has proved him right, and he has said so, pointing chiefly at the Casarabe centres of Cotoca and Landívar in Bolivia.

The wider claim is where the two accounts separate, and Upano separates them unusually cleanly. Hancock's Amazonian cities are meant to be a trace of something older and external: survivors of a lost Ice Age civilisation seeding agriculture and monumental building into the Americas. What the Upano sequence actually shows is the opposite of an arrival. The platforms accrete over centuries, the ceramic styles change gradually from Kilamope to Upano to Huapula, the labour is well within the reach of a coordinated agrarian population, and the material culture is unmistakably indigenous and local. There is no intrusive layer, no imported technology, and nothing that requires anyone to have come from anywhere.

A quieter argument concerns the language rather than the past. Archaeologists have grown wary of the 'lost city' framing that carried Upano through the world's newsrooms — a study in Advances in Archaeological Practice examined precisely how LiDAR results get presented as lost-world discoveries, warning that the dramatisation erodes the public's ability to tell good science from bad. Nothing here was lost. The Shuar have lived in this valley continuously; the settlements were abandoned, not vanished; and describing an indigenous achievement as a mystery solved by lasers quietly writes its builders out of their own story.

Key evidence cited

The discovery decisively confirms the heterodox claim that Amazonian forest concealed substantial ancient urbanism
Early Spanish chronicles describing dense populations in the upper Amazon look far more credible in its light
Hancock cites the Casarabe centres of Cotoca and Landívar as vindication of the Amazonian argument in America Before
Vast tracts of the basin remain unsurveyed, so arguments from the present absence of evidence are demonstrably weak
The episode shows an entrenched expert consensus — Meggers's 'counterfeit paradise' — can be comprehensively overturned

Genuinely open questions

How many people actually lived in the valley at its peak, given how weakly the platform count alone constrains population?
Did Sangay's eruptions end the system, and can tephra layers be tied tightly enough to the abandonment to prove it?
Is 'city' the right word for low-density garden urbanism, or does borrowing it import expectations the evidence cannot meet?
How far beyond the surveyed 300 km2 does the network extend, and what would full-valley LiDAR coverage add?

Worth knowing

Rostain had been digging at Sangay and Kilamope for a quarter of a century without grasping what he was standing in. The valley only became a city once the trees were deleted from the data — the discovery was made not in the field but on a screen, from a point cloud flown in 2015 and not fully published for another nine years.