A study published July 29, 2026 in the journal Nature identifies more than 20,000 precolonial earthworks across the southwest Amazon, using lidar (light detection and ranging) to pierce through dense rainforest canopy and document structures that had eluded ground-based survey for decades.

The research team classified the majority of the newly discovered structures as geoglyphs — large-scale earthen formations laid out in geometric patterns — concentrated in the border region of Brazil, Bolivia, and potentially adjacent countries, according to reporting from Live Science. Science confirmed that lidar was the primary detection technology and that the team identified most of the earthworks as geoglyphs.

The scale of the find dwarfs prior estimates. A 2023 study by V. Peripato et al., published in Science, had estimated that more than 10,000 pre-Columbian earthworks remained hidden across the Amazon basin — a figure that the new Nature study now appears to have substantially exceeded through systematic remote sensing. Before that, a 2017 PNAS study by J. Watling et al. had documented over 450 geometric ditched enclosures across roughly 13,000 km² of Acre state, Brazil, offering an early quantitative baseline for geoglyph density in the region. The trajectory from 450 structures in 2017 to over 20,000 in 2026 reflects not just incremental discovery but a methodological shift: lidar enables coverage of terrain that is either impenetrable on foot or has been degraded by deforestation before archaeologists can reach it.

The Max Planck Institute for Geoanthropology (MPI-GEA) had signaled the direction of this research in an October 2023 press release titled "Amazonian Rainforest Hides Thousands of Records of Ancient Indigenous Communities Under Its Forest Canopy." A subsequent Mongabay report in December 2023 noted that researchers had identified 24 previously unreported pre-Columbian earthworks in the Amazon, connecting the archaeological findings to contemporary Indigenous land struggles. Earlier this year, Daniel Pierce and Christopher Bodine used lidar data to uncover geoglyphs deep in the Amazon, as reported in February 2026, adding to a growing body of independent and institutional efforts deploying remote sensing in the region.

The broader context here is a sustained reevaluation of precolonial Amazonian demography and social complexity. For most of the twentieth century, the prevailing orthodoxy held that the Amazon's nutrient-poor soils (particularly oxisols and ultisols) could not support the dense, sedentary populations necessary for monumental earthwork construction. The successive geoglyph surveys — from Watling's 450 enclosures in Acre to the current inventory of over 20,000 structures — have steadily eroded that model. The sheer density of earthworks in the southwest Amazon implies landscape modification at a scale consistent with substantial, organized, and possibly hierarchical societies operating over long time horizons.

The geoglyphs themselves remain functionally debated. Interpretations range from ceremonial or ritual centers to defensive enclosures, water-management infrastructure, and territorial markers. The new Nature dataset, given its scope, may provide the statistical power to test competing hypotheses about geoglyph function by examining spatial relationships between sites, their proximity to water sources, and variations in geometric form across sub-regions.

There are also conservation stakes. Many earthworks lie in actively deforested zones; some are likely destroyed before documentation. The lidar record, in this sense, functions as a preservation instrument: it captures an archaeological baseline that ground conditions may not maintain. The overlap between ancient earthwork distributions and contemporary Indigenous territories, as Mongabay reported, also situates these findings within ongoing disputes over land tenure and resource rights in the Amazon basin.

The study does not resolve questions about which specific precolonial populations built the earthworks, over what chronological span, or through what labor organization. What it does is establish an empirical floor — over 20,000 structures, mapped at scale — that any future reconstruction of Amazonian precolonial history must account for.