AI Tool Unlocks the Ancient Secrets of Sand Grains

A new AI tool, SandAI, developed by Stanford researchers, unveils the ancient history of quartz sand grains, providing insights for both geological studies and modern-day forensic investigations

SEM image of a sand grain — Scanning electron microscopy reveals the shape and texture of a quartz sand grain from the Mississippi River. The pictured sand grain is about 200 micrometers in length. *Credit: Michael Hasson*

Researchers at Stanford University have developed a cutting-edge artificial intelligence tool, called SandAI, that can trace the geological history of quartz sand grains. This innovative technology offers insight into the processes that shaped sand, whether through wind, rivers, waves, or glaciers, across hundreds of millions of years.

SandAI brings unprecedented precision to geological and archaeological studies, helping researchers decode the past in areas where other evidence, such as fossils, is sparse. The tool’s advanced microtextural analysis could also prove useful in modern-day forensic investigations, such as combatting illegal sand mining.

“Working on sedimentary deposits that haven’t been disturbed or deformed feels about as close as you can get to being in a time machine – you’re seeing exactly what was on the surface of Earth, even hundreds of millions of years ago,” said Michael Hasson, a PhD candidate at Stanford and lead author of the study published in Proceedings of the National Academy of Sciences. SandAI, developed in collaboration with Earth and planetary sciences assistant professor Mathieu Lapôtre, is already transforming how researchers study sedimentary deposits.

From Ancient Clues to Modern Technology

Traditionally, microtextural analysis has involved manual observation through microscopes, where scientists could make educated guesses about a grain’s transport history. Modern research confirmed that these processes—such as river flow or wind—leave specific marks on sand grains. However, manual analysis is often subjective, slow, and inconsistent.

“Instead of a human going through and deciding what one texture is versus another for sand grains, we are using machine learning to make microtextural analysis more objective and rigorous,” said Lapôtre. The application of this AI-driven method expands the scope of research far beyond its previous limits.

Training the AI for Global Insight

Hasson and his team trained SandAI using hundreds of scanning electron microscope images of sand grains from a variety of environments, including riverbeds, deserts, glaciers, and beaches. The goal was to create a comprehensive tool capable of working across geological time and geography. SandAI has since achieved an impressive 90% accuracy in predicting the history of sand grains, including samples from as far back as 200 million years.

The team further tested the tool on sand grains from the Cryogenian period, over 600 million years ago. These grains were likely part of ancient windblown dunes near glaciers, as evidenced by similar findings from modern Antarctic sand samples. “These findings from SandAI suggest that Antarctica really is a good modern analog to the environment represented by the Bråvika Member,” said Hasson. SandAI’s ability to cross-reference ancient and modern data is paving the way for new discoveries about Earth’s ancient environments.

Modern Applications of Ancient Science

Beyond its academic value, SandAI is being hailed for its practical applications in addressing environmental challenges. Sand, the second most-used resource in the world after water, is critical for the construction industry. SandAI’s ability to trace the origins of sand could help ensure ethical sourcing, as well as aid in tracking illegal sand mining.

“The fact that we can now offer detailed conclusions about geological deposits that weren’t knowable before, I find kind of mind-blowing,” said Hasson. The research team has made SandAI freely available online, encouraging other scientists to explore its potential.

SandAI’s versatility may yet unlock more secrets buried in the sand, offering new avenues of research and potential solutions to modern-day issues.

You can read the full article “Automated determination of transport and depositional environments in sand and sandstones” in Proceedings of the National Academy of Sciences,