The Amazon Was Not Empty. LiDAR Has Made That Harder to Pretend.

TL:DR

• A 2024 airborne laser survey across 4,497 km² of forest in Acre and Amazonas, Brazil, recorded 432 earthworks; 396 had not previously been documented.
• The directly observed result is substantial. The headline estimate—24,000–30,000 earthworks across the wider Aquiry cultural area, and 1.25–3 million people at its peak—is a modelled extrapolation, not a count.
• The structures include geometric ditched enclosures, embankments and roads. The evidence best supports recurring public, ceremonial and political centres; it does not establish a single empire, a “city” under every tree, or an exact population.
• The immediate consequence is methodological: canopy-penetrating LiDAR has exposed a serious detection bias in archaeology based mainly on deforested land and satellite imagery.

Four hundred shapes beneath the canopy

The decisive number is not three million. It is 396.

That is the number of previously unrecorded earthworks identified when Martti Pärssinen and colleagues flew airborne laser-scanning transects across southwestern Amazonia in 2024. Their aircraft travelled 4,430 km along ten survey lines over the Brazilian states of Acre and Amazonas. From laser returns that reached the ground through gaps in the canopy, the team built terrain models and documented 432 earthworks, only 36 of which were already known.

Some are geometric ditches and embankments—squares, circles, polygons, sometimes linked by straight roads. The ditches commonly measure 9–13 metres wide and can reach 5 metres deep; individual geoglyphs range from 0.35 to 14 hectares, while the largest known example approaches 50 hectares.

The research team uses Aquiry for a multicultural complex of communities associated with these architectural and ceremonial traditions. It is a modern archaeological designation, derived from an Indigenous name for the Acre River—not the recovered self-name of a single, centrally governed people.

That distinction matters. The forest has not yielded an Amazonian version of Rome waiting intact beneath leaves. It has yielded a denser record of long-lived, organised landscape-making than the old map could see.

The finding, and the leap beyond it

The observed finding is strong: LiDAR detected hundreds of earthworks in a defined survey area. The wider claims are plausible but less settled.

Using the LiDAR results alongside earlier satellite and aerial-image surveys, the authors estimate that the Aquiry cultural area covered roughly 183,000 km² and may contain 24,000–30,000 earthworks. They propose a peak population of 1.25–3 million people around AD 100–300.

This is where responsible reading begins. A LiDAR detection is not the same thing as an excavated, dated and culturally attributed site. And an estimate of sites is not an exact census of inhabitants.

Independent archaeologists quoted by Scientific American welcomed the study’s corrective to the “untouched wilderness” idea but flagged two live uncertainties: how nearby structures were grouped or separated as sites, and whether a high number of earthworks could partly reflect frequent relocations rather than consistently high population density. Michael Heckenberger, speaking to New Scientist, called the population claim plausible and important, while identifying the next research task: finding and characterising the occupational sites around the visible earthworks that would support the estimate.

That is not a flaw to hide. It is the actual research agenda.

What LiDAR changed

Remote sensing did not create these structures. It changed the sampling problem.

Before canopy-penetrating scans, discovery was biased toward cleared land, where earthworks appear in aerial photographs and satellite imagery. In forested country, vegetation erases their surface geometry from ordinary view. LiDAR sends pulses from an aircraft; the subset that reaches bare ground can be classified into a digital terrain model. In this study, that made ditches, banks and roads legible without excavation.

The result is a warning against treating what is easy to observe as a neutral record of what exists. The team found 156 earthworks in areas already examined with satellite imagery, where prior searches had identified only 30—a roughly fivefold difference in that comparison.

That is the durable scientific signal. Archaeological absence in rainforest is often an instrumentation problem before it is evidence of historical absence.

What this is not

It is not permission to replace one caricature with another.

The old caricature pictured the precolonial Amazon as pristine, thinly inhabited nature. The new viral caricature will picture it as a uniformly dense, fully engineered garden-city from border to border. The study supports neither.

Its evidence is concentrated in southwestern Amazonia. It finds pronounced variation, including flight lines where no earthworks were detected; the authors explicitly note that this could represent either a technical detection constraint or ancient buffer zones without permanent occupation. Later mound villages and rounded embanked enclosures also belong to different periods from the Aquiry geoglyph-building tradition.

The appropriate correction is sharper: parts of Amazonia supported large, socially complex populations that reshaped forests, soils and ceremonial landscapes over centuries. How far that pattern extends, how populations were distributed, and how ecological change varied from one region to another remain open questions.

A forest can be ancient and human-shaped at once

The non-obvious implication is ecological, not merely archaeological.

The study connects the earthwork record to prior evidence of cultivation, forest burning, horticulture and management of useful plants. That makes the Amazon’s present ecology partly a cultural archive: a forest can be biodiverse, old and still bear the long imprint of human choices.

This does not establish that past land use was uniformly sustainable, nor does it turn archaeology into a ready-made prescription for contemporary conservation or climate policy. It does mean that biodiversity baselines built on a fantasy of wholly people-free rainforest are historically incomplete.

The most important next work will pair broad LiDAR coverage with excavations, radiocarbon dating, palaeobotany, soil studies, and the leadership and knowledge of Indigenous communities whose territories and histories are implicated. The map is now ahead of the explanation.

Stakeholder landscape

• Indigenous peoples and local communities: The findings concern living cultural landscapes, not empty research terrain. Site protection, consent and benefits from future research matter as much as discovery counts.
• Archaeologists: LiDAR can set priorities for fieldwork, but it cannot substitute for ground verification, dating or interpretation.
• Conservation planners: Forest protection has a heritage dimension; clearing can destroy both habitat and archaeological context.
• Climate and ecology researchers: Models of forest composition, carbon and “natural” baselines may need to account more carefully for deep human landscape modification—if comparable evidence accumulates elsewhere.
• The public: The useful update is intellectual, not touristic. Do not seek out, publicise or geolocate vulnerable sites.

What this means for you

For readers: Update the mental picture, not the facts beyond their evidence. The study directly documents a much denser earthwork record in one southwest-Amazonian region. Treat continent-wide population claims as a question under active investigation.

For educators and communicators: Use “Aquiry cultural area” or “communities associated with Aquiry traditions,” rather than implying a monolithic lost empire. Lead with the measured count—432 features, 396 newly documented—then label the 24,000–30,000 and 1.25–3 million figures as estimates.

For people sharing this story: Avoid publishing precise coordinates or framing the discovery as an invitation to visit. The responsible action is support for Indigenous rights, heritage protection and forest conservation, not amateur site-hunting.

Uncertainty ledger

Bottom Line

Airborne LiDAR has not found a single lost city. It has revealed a detection failure in the way the Amazon’s past has been mapped—and, in doing so, added 396 previously unseen earthworks to the record. The study makes large precolonial populations in southwestern Amazonia more credible; it does not turn its extrapolated population range into settled fact. The real discovery is a more demanding one: the rainforest is not only a biological inheritance. In many places, it is also a human historical landscape.

Written in the tradition of — E.

Sources

• Tier 1 — Primary research: Pärssinen, M. et al. “Over 20,000 precolonial earthworks in Southwest Amazonia.” Nature (29 July 2026)
• Tier 2 — Independent science reporting: Jackie Flynn Mogensen, “Hundreds of ancient, never-before-seen ‘earthworks’ discovered in the Amazon.” Scientific American (29 July 2026)
• Tier 2 — Independent science reporting: James Woodford, “Lost civilisation in the Amazon may have numbered 3 million people.” New Scientist (29 July 2026)
• Tier 2 — Independent science reporting: Erik Stokstad, “Odd shapes hidden in dense Amazon rainforest reveal sprawling ancient civilization.” Science (29 July 2026)