Scientists were able to predict a volcanic eruption using a supercomputer. Article about it published in Science Advances.
Geologist Patricia Gregg of the University of Illinois at Urbana-Champaign and her colleagues created a program in 2017 to model volcanic processes. It was designed to run on Blue Waters and iForge supercomputers. The model gained attention after she managed to recreate the unexpected eruption of Okmok volcano in Alaska in 2008.
Using this model, the scientists studied the volcano Santo Tomas in Ecuador, which may erupt in the near future. “Santo Tomas is a benign volcano. That means it shows all the signs of an eruption, such as gas emissions and seismic activity. This made it a good test subject for our model,” Gregg said.
In January 2018, scientists ran a simulation that showed that between June 25 and July 5, the magma chamber would become unstable, possibly leading to an eruption. As a result, the eruption occurred one day after the earliest predicted date.
The advantage of this model lies in its ability to continuously absorb data from different sources in real time and process it quickly to provide a daily forecast similar to a weather forecast.
Predicting eruptions is one of the most difficult tasks in volcanology, which will help to prepare for them in the future and prevent human losses.