AI SCIENCE
AI just exposed Yellowstone’s underground drama
Scientists have used machine learning to scan 15 years of seismic data from Yellowstone and uncovered over 86,000 previously undetected earthquakes.
The findings offer a whole new view of the volcanic system simmering beneath one of America’s most famous national parks.
Yellowstone isn’t just geysers and hot springs; it sits on one of the most active volcanic systems on the planet.
By running years of seismic records through an AI model, researchers found earthquake “swarms”: clusters of small, fast-moving quakes concentrated along immature, rugged fault lines that were missed in earlier analysis.
Why this matters:
AI detected 10x more earthquakes than manual methods.
Over half were part of fast-moving, tightly packed swarms.
Findings could improve safety forecasts and support energy planning.
Seismic tea
These swarms were found to move through rougher fault zones and seem to be triggered by a mix of slow-moving underground water and sudden fluid bursts.
The study also used fractal modelling, think broccoli or snowflakes, to understand the irregularity of these patterns and why they behave the way they do.
The research, published in Science Advances, shows how AI is changing the way scientists study earthquakes.
It’s not just about Yellowstone, this approach could help make risk forecasting more accurate and help guide geothermal energy projects in other active areas around the world.
Me: "It's just hot springs." Yellowstone: unleashes 86,000 secret earthquakes.