Researchers using artificial intelligence and satellite images have identified traces linked to ancient forests in the Atacama Desert, offering a new way to study landscapes that once supported human communities in one of the driest regions on Earth.
The study, led by Elena Balmaceda-Baraona, focused on the Pampa del Tamarugal in northern Chile. The Atacama is widely known for its extreme dryness and harsh conditions, which can make large-scale field research difficult. Researchers used satellite data and machine learning to detect spectral signals associated with former forest patches. The method does not directly show vanished forests. Instead, it identifies surface characteristics linked to surviving botanical remains, including wood, leaf litter, charcoal, and tree imprints.
Satellite images reveal hidden landscape patterns
Researchers analyzed free multispectral images from the European Space Agency’s Sentinel-2 satellites. Multispectral sensors record reflected light across several wavelengths, allowing scientists to distinguish materials that may look similar in ordinary photographs.
The team combined those images with information collected in the field. Researchers then used supervised classification, a machine learning method trained with known examples, to identify areas with similar spectral patterns. Field observations helped test whether the satellite results matched locations containing evidence of former forests.
The approach follows a growing trend in which archaeologists use AI and remote sensing to investigate difficult desert landscapes. Such methods can help researchers examine large areas before deciding where to concentrate fieldwork.
Ancient forests mattered to Atacama communities
The forests are important because people have occupied the region for roughly 12,000 calibrated years before present. Previous archaeological research has shown that the Pampa del Tamarugal’s forests became part of local subsistence practices and cultural traditions. Wooded areas provided resources in an otherwise extremely dry environment and also became connected with the beliefs of communities living there.
AI and satellite images are revealing traces of ancient forests hidden across Chile’s Atacama Desert.
Researchers detected signs of wood, charcoal, leaf litter and tree imprints, offering new clues about landscapes used by people thousands of years ago. pic.twitter.com/WwUGXztV4l
— Tom Marvolo Riddle (@tom_riddle2025) October 1, 2026
Researchers compared the newly detected forest proxies with the locations of known archaeological sites. The analysis could help clarify how ancient communities interacted with changing environments across the desert landscape. Archaeological discoveries elsewhere in the Atacama have also provided evidence of how people adapted to the region. Researchers recently identified 5,000-year-old spear-throwers from hunter-gatherers along Chile’s Atacama coast.
AI could guide future archaeological surveys
Machine learning is increasingly allowing archaeologists to analyze large landscapes and identify patterns that can be difficult to recognize from the ground. One striking example comes from Peru, where researchers used AI to help identify hundreds of previously unknown Nazca geoglyphs. The Atacama study takes a different approach. Instead of searching directly for human-made structures, it focuses on environmental evidence associated with landscapes used by ancient communities.
Researchers describe the model as exploratory. It cannot replace field surveys or excavation. However, it could help archaeological teams identify areas that deserve closer investigation before beginning work on the ground. The findings show how satellite imagery, machine learning, and traditional archaeology can work together to reconstruct parts of ancient environments that are difficult to recognize from the desert surface.
See all the latest news from Greece and the world at Greekreporter.com. Contact our newsroom to report an update or send your story, photos and videos. Follow GR on Google News and subscribe here to our daily email!
