Spanning Wyoming, Idaho and Montana, the Yellowstone Caldera is one of the most seismically active volcanic regions on the planet. When the volcano erupts, a caldera forms, emptying the magmatic chambers below, and the above land collapses into a major depression.
In a recent groundbreaking study, Professor Bing Li, a Canadian Western University and his team, collaborated with researchers at De Santander University in Columbia and the US Geological Survey, used machine learning to go through 15 years of historic seismic data from Yellowstone.
Their advanced methods, which were sized and assigned to about ten times more seismic events than previously recorded, provided new insights into ongoing dynamics under the Yellowstone Caldera and improved their ability to assess potential volcanic risks.
Over half of Yellowstone earthquakes occur in groups
The new study significantly expands the Yellowstone Caldera seismic record. This includes 86,276 events from 2008 to 2022. This enhanced catalogue provides much clearer images of the volcanoes and seismic activity in the region thanks to improved data analysis and system reviews.
One important finding from this study is that more than half of these earthquakes occurred as part of a herd of earthquakes. This is a small cluster of linked tremors that moves and evolves within a confined area for a short period of time. This insight will help scientists better understand the complex behavior of Yellowstone's underground systems.
According to Professor Li, earthquake swarms are different from aftershocks. They consist of small, interconnected earthquake clusters occurring in short periods and within a limited area, rather than a sequence of tremors following a single, larger event. Li emphasizes that although each volcano, including Yellowstone, has its own unique characteristics, the patterns observed in these herds provide valuable insights.
Understanding these seismic patterns can also help to improve safety protocols, provide clearer risk information to the public, and guide the development of geothermal energy by avoiding dangerous zones with heat flow, Li pointed out.
AI helps to reveal hidden earthquakes in historical data
Before machine learning could be used, experts had to manually inspect seismic data, which was slow, expensive and missed many small events. Today, machine learning allows scientists to analyze large amounts of historical earthquake data stored in databases around the world.
The new approach reveals more earthquakes and helps researchers better understand both famous and hidden seismic areas around the world. Professor Li points out that manually analyzing huge amounts of seismic data is simply feasible but not scalable.
In the latest research, earthquake swarms beneath the Yellowstone Caldera occur along relatively young and rough fault structures. This differs from the more developed and smoother faults found in Southern California and areas just outside the caldera.
The researchers emphasized that they are equipped with a detailed and reliable catalogue of seismic activity under the Yellowstone Caldera, and that they apply advanced statistical methods to identify and examine previously undetected earthquake swarms, allowing for a deeper understanding.
This study is published in the journal Advances in science.
