Geoglyphs of Acre

← Encyclopedia

When the cattle came to eastern Acre in the 1970s and 1980s, the chainsaws and the bulldozed pasture did something no archaeologist had been able to do: they took the roof off. Where humid upland rainforest had stood unbroken — the kind of canopy that tropical ecologists of the period read as primeval, scarcely touched by human hands — the cleared ground showed a geometry too exact to be accidental. Perfect circles two and three hundred meters across. Squares with squared corners. Rings set inside squares, double-ditched figures, the occasional three-sided U. From the air, once the trees were gone, the patterns leaped out; from the ground, standing in a ditch overgrown for centuries, a person could walk the whole figure and never see its shape. The forest had not hidden a ruin. It had hidden a design.

These are the earthworks now cataloged, somewhat misleadingly, as the geoglyphs of Acre — hundreds of monumental ditched enclosures cut into the firm interfluvial ground of the western Brazilian Amazon, concentrated in the eastern part of the state of Acre near Rio Branco, between the valleys of the Acre and Iquiri rivers. More than 450 such sites have been recorded in Acre alone, distributed across roughly 13,000 square kilometers of what was, until the lifetime of people now living, continuous rainforest. The wider survey across southwestern Amazonia — taking in adjacent Rondônia and northern Bolivia — has pushed the count of earthwork structures into the thousands as satellite and deforestation surveys extend the mapped area year by year. The discovery has been called the most dramatic recent find in Amazonian archaeology, and the description is not inflated: it overturned, in a single generation of clearing, the assumption that this stretch of forest had ever been empty.

A geometry in the ground

The word geoglyph is a borrowing, and an imperfect one. It carries the image of the Nazca lines of Peru’s south coast — figures and biomorphs drawn on a desert surface to be read from above — and the Acre works are something else entirely. They are not drawings on a surface but excavations into it: geometric figures defined by a ditch and, beside it, the bank of earth thrown up from that ditch. A typical enclosure is a circle or a quadrangle on the order of 100 to 300 meters across, its perimeter a ditch up to roughly eleven meters wide and one to two meters deep, the spoil heaped along the outer edge into an embankment. Where ditch-bottom and bank-top survive together, the original drop from crest to floor was more than five meters — a serious earthen architecture, not a shallow scratch. Many enclosures interlock: a ring tangent to a square, a figure nested within a figure, sites where two or three distinct shapes share ground within a few hundred meters. Linear features — raised causeways and sunken roads, walled avenues running out from the openings — connect some enclosures to others and to the streams and springs along the plateau edges.

Each enclosure has openings, gaps left in the ditch-and-bank ring, and the openings are not random; at many sites they align consistently, hinting at orientations laid out by intention rather than convenience. To build one demanded the controlled removal of an enormous volume of earth by hand, the spoil shaped and placed with care, the whole figure set out true on the ground without any vantage from which its plan could be seen. That alone tells of organized labor and a shared template — a knowledge of form held in common and executed across a landscape, not improvised hut by hut.

The distribution, too, is patterned rather than scattered. The enclosures favor the flat, well-drained interfluve — the firm ground of the plateaus that does not flood — set back from the major rivers and clustered near the headwater streams and springs that fed them. They keep a recurring repertoire of shapes: the single ring, the single square, the ring-within-square, the paired or double-ditched figure, with composite sites where several forms stand within walking distance and roads tie them together. The spatial analysis of the set finds regularities in spacing and in the recurrence of particular geometries that argue for a regional convention — a way of marking ground shared across the eastern Acre landscape, not a series of unrelated local experiments. Whoever cut these figures carried the same idea of how a place of consequence ought to be shaped, and reproduced it again and again across a territory the size of a small country.

The problem of the dates

Dating the Acre works is genuinely hard, and the published ranges are wide enough to be honest about the difficulty. The first scientific notice goes back to 1977, when Ondemar Dias recorded a handful of earth structures during the National Program of Archaeological Research of the Amazon Basin; the sites drew little attention until the early 2000s, when Alceu Ranzi of the Federal University of Acre and Martti Pärssinen of the University of Helsinki assembled an international team, soon joined by the Brazilian archaeologist Denise Pahl Schaan, and began systematic survey and excavation.

Radiocarbon assays on charcoal from the ditches and banks place the main floruit of enclosure-building roughly between 2,000 and 650 years ago — most activity falling between about 100 and 400 CE, with a marked decline after 1000 CE. Individual sites carry firm anchors: a charcoal sample from Fazenda Colorada returned an age that calibrates to 1244–1378 CE; the oldest road sample at the Tequinho site, tied to that enclosure’s northern avenue, dates to between 63 BCE and 124 CE. Some figures appear to have been initiated as early as the mid-first-millennium BCE and to have been reused — even, on the most recent readings, reoccupied or recut as late as the nineteenth century, long after the building tradition itself had lapsed. The dating problem is compounded by the way the enclosures were used: a ditch dug once and kept open for centuries accumulates charcoal across its whole history, so a single date fixes a moment of activity, not the act of construction. The chronology is best read as a long arc — centuries of building, centuries of use, slow abandonment — rather than a single founding event.

What were they for?

On function there is no settled consensus, and the entry is worth holding open as a real debate rather than a solved case. Three readings have been argued. The first treats the enclosures as defensive — ditch-and-bank as fortification, the classic interpretation of earthworks the world over. The second treats them as domestic, the footprints of villages. The third, which has steadily gained the most ground, treats them as ceremonial: monumental public spaces where dispersed communities gathered, the Amazonian counterpart of the civic-ceremonial centers of the Andean Formative.

The decisive evidence pushes against fortification and against ordinary settlement alike. Excavation inside the enclosures turns up remarkably little: thin scatters of pottery, hearths, the residues of feasting, but not the deep, dense midden of a place lived in continuously. The ditches are often placed so that the bank lies outside them — the reverse of a defensive arrangement, where the wall should stand behind the obstacle. And the consistent orientation of the openings, the formality of the geometry, the linked roads and avenues, all read more naturally as the architecture of assembly and procession than of dwelling or defense. Denise Schaan and her colleagues set the geoglyphs within a sacred geography: places built for periodic congregation, ritual, and the marking of a worked landscape, used by people who lived elsewhere and came together here. The Tequinho complex, with its radiating roads, has been read in just this way — a central node for gathering, its avenues both ceremonial routes and boundaries inscribed across the ground.

This is not to make the enclosures otherworldly. The phytolith record from Tequinho and the JK site preserves maize and squash, evidence that the cleared ground around the figures was also farmed; the line between the sacred precinct and the cultivated patch was never sharp. The ethnic affiliation of the builders has not been firmly established. Some recent work points toward a long Arawak and Takana presence in the region — the ancestors, on one reading, of peoples such as the Manchineri and Apurinã — but the attribution is provisional, and the works are not assigned with confidence to any single living group.

A forest that was made

The most consequential thing the geoglyphs revealed lay not in the ditches but in the soil around them. The reigning twentieth-century model held the western Amazon too poor in soil to have ever supported large, settled, landscape-shaping populations — fit only for thin bands of shifting cultivators. The enclosures are the wrong size and the wrong number for that picture: they imply coordinated labor, durable social organization, and a population dense enough to raise monuments across thirteen thousand square kilometers.

The paleobotanical work sharpened the point. Phytolith and charcoal cores taken at the enclosures and dated against the local sequence show that this was bamboo-dominated forest for at least 6,000 years — a vegetation resilient to both climate swing and human use. Crucially, the geoglyphs were not hacked from virgin jungle. The builders cut only small, temporary clearings to lay out and use the figures, and they did so within a forest that human hands had already been shaping for millennia. Beginning roughly 4,000 years ago, small burned openings were used to draw in economically valuable trees; pre-geoglyph clearance episodes around 4,000 and 3,600 years before present were followed by a sustained, multi-millennial enrichment of palms — a deliberate cultivation of the forest into a standing larder. The forest that grew back over the abandoned enclosures still carries that signature: stands unusually rich in palms and in Brazil nut (Bertholletia excelsa), a botanical legacy of the people who built here.

This is the larger claim the Acre earthworks press: that pre-Columbian forest societies did not merely inhabit the Amazon but engineered it — clearing, planting, mounding, ditching, and routing water across the land — so that much of what later eyes took for pristine wilderness is, in fact, an old human work grown over, kin to the ditched enclosures of the Casarabe plain to the south. The archaeologists who established this attach a caution to it that belongs in the record: the deep history of managed forest is no license for the destructive clearing of the present, and the slow, enriching land-use that maintained a largely forested Acre into the mid-twentieth century is the opposite of what the bulldozers brought.

The comparison with Nazca repays care precisely because it must not be collapsed. Both are monumental ground-marking; both reorganize a landscape at a scale legible from above. But Nazca’s figures are surface lines scraped into a hyper-arid desert pavement and meant to be walked or seen, while Acre’s are deep ditched enclosures cut into living rainforest — a different technique, in a different climate, raised by a different people a continent’s reach away, and across a span of centuries that barely overlaps. The shared term “geoglyph” hides more than it captures: it groups two distinct technologies of marking the earth under one borrowed name, and the borrowing has done as much to confuse the Acre works as to describe them.

The textual and scholarly record

The Acre earthworks are an actively contested research field, and the literature divides along the two questions the sites raise — what the enclosures were, and what they say about the forest. On the landscape question, the key paper is Jennifer Watling, José Iriarte, Denise Schaan and colleagues, “Impact of pre-Columbian geoglyph builders on Amazonian forests,” Proceedings of the National Academy of Sciences 114 (2017), which established from phytolith and charcoal cores that the builders worked in already-managed forest rather than pristine jungle. Martti Pärssinen and colleagues extended the chronology deep into the Holocene in “The geoglyph sites of Acre, Brazil: 10,000-year-old land-use practices and climate change in Amazonia,” Antiquity 95 (2021), arguing that human land-use in the region reaches back millennia before the first ditch was cut. Denise Pahl Schaan’s Sacred Geographies of Ancient Amazonia: Historical Ecology of Social Complexity (Routledge/Left Coast Press, 2012) is the synthetic monograph that frames the enclosures as ceremonial sacred geography within the wider Amazonian record. The function debate and the early excavations are documented in Schaan, Pärssinen and colleagues’ site reports, including the Fazenda Atlântica and Quinauá work in Latin American Antiquity, and the broader land-use case is summarized for a general readership by Agência FAPESP. The cluster’s place on Brazil’s heritage agenda is registered in the UNESCO World Heritage Tentative List entry for the Geoglyphs of Acre. Lidar survey and the analysis of associated roads and mound villages — the Tequinho road complex among them — have refined the settlement picture, distinguishing the residential mound villages of southeastern Acre from the ceremonial enclosures to which their inhabitants came.

Every dry season the deforestation front advances, and every advance peels another stretch of canopy off the plateau. The pattern repeats with each clearing: where the trees fall, another ring or square surfaces, true-cornered and centuries-silent, and the count climbs again. What is emerging from under the forest is not a scatter of ruins but a single engineered ground — drained, routed, planted, and inscribed with geometry — an archive of design that the canopy preserved by hiding, and that the chainsaw, against every intention, has begun to read back.

→ Related: Nazca · Casarabe Culture · Marajoara Culture · Santarem Tapajos Culture · Upper Xingu Kuhikugu