Study finds 10 times more quakes in Yellowstone than earlier records showed
Machine learning algorithms applied to waveform data from 2008 to 2022 have revealed 86 276 earthquakes beneath the Yellowstone caldera, U.S., approximately 10 times more than previously recorded. The revised catalogue, published in Science Advances on July 18, 2025, was created by researchers from Western University, Universidad Industrial de Santander, and the U.S. Geological Survey.
The Grand Prismatic hot spring in Yellowstone National Park is sourced from a magma chamber beneath it. The bright colours are produced by hydrophilic bacteria in the mineral-rich water. Credit: Bing Li
A newly compiled seismic catalogue based on 15 years of waveform data shows that the Yellowstone caldera, a large volcanic depression spanning parts of Wyoming, Idaho, and Montana, experienced 86 276 earthquakes between 2008 and 2022.
This marks a tenfold increase compared to the previously known number of events and was made possible through the application of advanced machine learning techniques and a region-specific 3D velocity model.
The study, published in Science Advances on July 18, was led by Bing Li of Western University in collaboration with Universidad Industrial de Santander (Colombia) and the United States Geological Survey (USGS).
It demonstrates how artificial intelligence can radically improve detection rates and characterization of microseismic activity in complex volcanic regions.
Prior to this effort, earthquake detection relied heavily on manual inspections and traditional algorithms, limiting the scale and granularity of the seismic record. To overcome these limitations, researchers trained a separate AI model for each seismic station in the Yellowstone network.
This approach allowed accurate magnitude assignment, even during periods of overlapping swarm events. In validation tests, the model recovered 83% of previously documented earthquakes and identified 855 new events over just a 10-day window, with over 99% of those confirmed as real earthquakes.
More than half of the earthquakes were found to occur in swarms, sequences of small earthquakes clustered in time and space, typically lacking a dominant mainshock.
These swarms migrated through immature, high-roughness fault segments rather than mature fault zones, unlike seismic regions such as southern California.
The roughness was quantified using fractal geometry, a method that measures the self-similarity and spatial irregularity of seismic activity.
The analysis revealed that swarms were likely triggered by a combination of slow fluid migration and sudden pressure changes in hydrothermal systems.
The underlying fault structures were significantly rougher than those outside the caldera, suggesting active deformation along developing fractures.
The machine learning framework also utilized a 3D seismic velocity model of Yellowstone’s crust.
This model helped accurately locate earthquakes and estimate magnitudes by accounting for heterogeneities in the subsurface that affect seismic wave propagation.
Researchers say the findings could help improve hazard assessments in other volcanic regions and support safer geothermal development. Better seismic imaging makes it easier to avoid areas where fluid movement often triggers earthquakes.
“By understanding patterns of seismicity, like earthquake swarms, we can improve safety measures, better inform the public about potential risks and even guide geothermal energy development away from danger in areas with promising heat flow,” said Bing Li.
References:
1 Machine learning reveals historical seismic events in Yellowstone caldera – Western News – July 18, 2025
2 Long-term dynamics of earthquake swarms in the Yellowstone caldera – Manuel A. Florez, Bing Q. Li et al. – Science Advances – July 18, 2025 – DOI: 10.1126/sciadv.adv6484 – OPEN ACCESS
Reet is a science journalist and researcher with a keen focus on extreme weather, space phenomena, and climate-related issues. With a strong foundation in astronomy and a history of environmental activism, she approaches every story with a sharp scientific lens and a deep sense of purpose. Driven by a lifelong love for writing, and a curiosity about the universe, Reet brings urgency and insight to some of the most important scientific developments of our time.
Commenting rules and guidelines
We value the thoughts and opinions of our readers and welcome healthy discussions on our website. In order to maintain a respectful and positive community, we ask that all commenters follow these rules.