Archaeologists working across three countries have identified networks of ancient settlements hidden beneath the Amazon rainforest, some spanning more than 100 hectares and connected by road systems stretching tens of kilometers. The findings, drawn from laser-scanning surveys in Ecuador, Bolivia, and Brazil, show that organized communities thrived in the Amazon basin as far back as 2,500 years ago, reaching a scale that researchers compare to medieval European towns. Taken together, the discoveries challenge a long-standing assumption that the Amazon was a sparsely populated wilderness before European contact.
Why ancient Amazonian urbanism reshapes modern conservation debates
The core tension behind these findings is practical, not just historical. Conservation policies across the Amazon basin have long treated the rainforest as pristine wilderness or, alternatively, as land available for agricultural conversion. Evidence of dense, long-lasting human settlement complicates both positions. If thousands of people once lived in planned communities across the Upper Amazon without triggering permanent deforestation, that record carries direct implications for how governments and international organizations think about sustainable land use in the region today.
These discoveries arrive as contemporary Amazonian nations face pressure to balance climate commitments with demands for mining, ranching, and industrial agriculture. For decades, the dominant narrative framed large-scale human activity as inherently incompatible with rainforest health. The newly documented settlement systems suggest a different story: pre-Columbian societies engineered landscapes, built roads, and managed water while maintaining extensive forest cover. That does not mean their practices can be simply copied into the present, but it does show that the forest is not a fragile, untouched ecosystem that only recently encountered human disturbance.
At the same time, archaeologists emphasize that historical land use does not justify current deforestation. Modern clearing often involves mechanized equipment, monoculture crops, and global commodity chains that operate on a scale and pace with no precedent in the archaeological record. The value of the new research lies in demonstrating a broader range of possible relationships between people and forest, expanding the menu of models that policymakers might consider when designing conservation frameworks and indigenous land rights regimes.
The Upano Valley data raises a specific, testable question about how these settlements functioned as a connected system. The road network mapped by LiDAR links clusters of earthen platforms across tens of kilometers in Ecuador’s Upper Amazon. If those settlements were occupied at the same time and connected by the straight causeways visible in the survey data, then residents of major centers could have reached neighboring sites within a few hours on foot. GIS cost-surface modeling, using the published platform coordinates and local terrain data, could estimate average daily travel times between centers and clarify whether these sites operated as an integrated urban network or as isolated communities that happened to share a road grid. That distinction matters because it determines whether the Upano Valley supported something closer to a regional economy or a loose scattering of villages.
LiDAR evidence from Ecuador, Bolivia, and Brazil
The strongest evidence comes from three separate research programs that used airborne LiDAR to see through the forest canopy. In Ecuador’s Upano Valley, a team led by archaeologist Stephen Rostain documented thousands of earthen platforms and plazas organized into clusters, with a regional road system extending tens of kilometers between settlement groups. The study, published in Science, describes the site as a 2,500-year-old example of what the authors call “garden urbanism,” where dense residential areas sat alongside managed agricultural land. The settlement pattern suggests a population large enough to build and maintain monumental earthworks over centuries.
In Bolivia’s Llanos de Mojos, a separate LiDAR survey published in Nature mapped Casarabe culture settlements that include two large sites each exceeding 100 hectares. Those sites feature monumental mounds and pyramids alongside engineered causeways, canals, and reservoirs. The infrastructure points to centralized planning and water management on a scale that required coordinated labor across a wide area.
A third line of evidence comes from the Upper Xingu region of the Brazilian Amazon, where researchers have documented clustered polities and engineered landscapes showing regionally integrated settlement patterns. The Xingu findings, published in Science, predate the LiDAR surveys and provided early scholarly arguments that parts of the Amazon supported organized, interconnected communities well before European arrival.
Outside the academic literature, reporting has highlighted how these discoveries overturn the image of a “virgin” forest, with accounts in outlets such as the Associated Press emphasizing that the newly revealed cities and roadways had been hiding in plain sight beneath dense vegetation. Those reports underscore how quickly the emerging archaeological consensus is filtering into broader public debates about the history and future of the Amazon.
What links these three sites across different countries and time periods is a shared pattern: planned settlements, engineered water and road infrastructure, and population densities that do not fit the older model of small, mobile groups scattered through unbroken forest. Each site was discovered independently, but together they form a body of evidence that is difficult to dismiss as a local anomaly.
Gaps in population data and the medieval comparison
Several questions remain open. The primary research papers from Science and Nature provide detailed counts of platforms, plazas, and road lengths, but they do not include tabulated population estimates for individual settlements. Population figures that appear in press coverage are inferred from platform density and site area rather than drawn from published demographic models in the peer-reviewed studies themselves. That gap matters because the headline comparison to medieval European towns depends on population scale, and the data to support a precise person-count is not yet in the published record.
The comparison to specific medieval European towns also lacks a formal citation trail in the primary literature. While the analogy is widely repeated, neither the Science nor the Nature papers include side-by-side data tables matching Amazonian sites to named European settlements of similar size. The comparison is useful as a rough frame of reference, but readers should understand it as an interpretive shorthand rather than a documented statistical equivalence.
Radiocarbon dating adds another layer of uncertainty. The published chronologies for these sites are aggregated across entire settlement complexes rather than tied to individual road segments or building phases. Without segment-level dates, it is not yet possible to confirm that all parts of a given road network were in use at the same time. If construction and abandonment happened in waves, the peak population at any single moment may have been smaller than the full footprint of platforms suggests.
The next development to watch is whether research teams apply GIS cost-surface analysis to the published Upano Valley coordinates. That work could establish whether the road system functioned as a tightly integrated transport grid or as a patchwork of connections linking only some communities at any given time. A clearer picture of contemporaneity would refine estimates of how many people the region could support simultaneously and how far political or ritual authority extended along the causeways.
Rethinking “pristine” forests and future research
Beyond population counts, the emerging picture of Amazonian urbanism raises broader questions about how archaeologists and ecologists define cities, towns, and rural landscapes. The settlements in Ecuador, Bolivia, and Brazil do not resemble the dense stone-built centers of the Andes or Mesoamerica. Instead, they appear as low-density mosaics of habitation, cultivation, and forest, blurring the line between urban and rural space. Some researchers argue that this pattern warrants its own category of urbanism, one that foregrounds ecological integration rather than hard boundaries between city and countryside.
Future work is likely to push in several directions at once. Expanded LiDAR coverage could reveal whether similar settlement systems existed in other parts of the basin that remain understudied. Targeted excavations on platforms, roads, and waterworks may clarify construction sequences and provide more precise dating. Paleoecological studies, including soil analysis and pollen records, could illuminate how ancient land use affected forest composition and biodiversity over the long term.
As those data accumulate, debates over conservation and development in the Amazon will increasingly have to grapple with a more complex human past. The new findings do not offer a simple template for present-day policy, but they do dismantle the binary choice between untouched wilderness and modern extractive frontier. Instead, they point toward a long history of negotiated landscapes in which people engineered their environments in ways that were intensive, organized, and, at least for centuries at a time, compatible with forest ecosystems.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
