'Volcanic tremors' can now be detected to forecast eruptions

It’s no secret that volcanoes are unpredictable, often hiding their secrets beneath layers of magma and age-old rock formations. However, an innovative solution to detect volcanic tremors is now in sight, thanks to relentless tenacity, and the power of technology.

The researcher at the helm, making sure we’re a step ahead in understanding these fascinating natural phenomena, is Darren Tan, a graduate student researcher at the University of Alaska Fairbanks Geophysical Institute.

Mapping volcanic tremors

Tan’s exciting breakthrough proposal comes down to this – an automated system that can document and classify the ongoing vibrations in active volcanoes. This system promises to entirely overhaul the long hours currently spent on manual documentation.

The innovative part? It’s built with the might of machine learning – a fascinating domain of artificial intelligence that leans on data to discern patterns and make decisions with the least possible human intervention.

This system will focus on volcanic tremor, a rhythmic seismic signal that escapes from volcanoes and often indicates the underground movement of magma or gas.

Volcanic tremor: Earth’s undercover rhythm

What’s volcanic tremor? It’s not a volcanic earthquake, even though both might seem similar. The volcanic tremor is a sustained ground rumble that can last anywhere from a few seconds to an entire year or even more. It’s so subtle that it’s usually detected in spectrograms because of its ever-changing intensity and frequency.

“Volcanic tremor isn’t typically detected or cataloged because it tends to be quite subtle in the seismic data. It doesn’t have the impulsive onset like an earthquake does,” explained Tan.

Currently, the Alaska Volcano Observatory, where Tan is affiliated, has manual processes to detect tremor. The seismologists there spend their days scanning spectrograms at 32 volcano-monitoring networks across Alaska to find even the faintest hint of tremor amidst the clear seismic signals.

Monitoring volcanic tremors

With 54 volcanoes classified as “historically active” in Alaska alone, such manual efforts can be time-consuming and tedious.

However, with Tan’s automatic system, this would be a thing of the past. He used the varied tremor signals from the 2021-2022 eruption of Pavlof Volcano to build a comprehensive dataset of labeled seismic and low-frequency acoustic spectrograms.

The result? Computer models that can detect and classify volcanic tremor in almost real-time. But don’t worry, human intervention will still be required to interpret what the automation churns out.

Shaking things up

The research paves the way for a new chapter in volcanic monitoring. “To be able to place our focus on time periods of interest, that is key,” said Tan. “I think that reinvents the way we can monitor long-duration eruptions, because things can get missed when a volcano is active for a year and a half or two years.”

The automated method doesn’t only detect tremor; it also contributes to forecasting and detecting eruptions, revolutionizing our approach towards these unpredictable giants.

Automation for seismic research

The implications of Tan’s automated system extend far beyond mere efficiency. By significantly reducing the time required for manual detection, researchers can now allocate more resources to analyzing the data and predicting volcanic behaviors.

This paradigm shift in seismic research could result in enhanced risk assessment methodologies, enabling better preparedness for potential eruptions.

Moreover, the integration of machine learning can open up new avenues for collaborative research, allowing data from various regions to be shared and compared more easily, leading to a global understanding of volcanic activity.

Future prospects

Looking ahead, the application of Tan’s system could set a precedent for the monitoring of not only volcanic tremor but also other geological phenomena.

As technology continues to evolve, similar approaches could be used to study earthquakes, landslides, and other natural events, utilizing machine learning and vast datasets to uncover patterns that were previously elusive.

By bridging the gap between advanced technology and the often unpredictable forces of nature, researchers can enhance our ability to predict and respond to environmental changes, ultimately safeguarding communities against geological hazards.

The Wild West of machine learning

Tan’s enthusiasm for machine learning is contagious. He dubs it the “Wild West of machine learning,” where everyone wants to dip their toes in.

But he warns that it’s essential to do so carefully. After all, we’re making decisions that will impact global safety responses and scientific understandings.

With Tan’s groundbreaking innovation, the future of volcanic monitoring is surely looking brighter. It’s not about predicting the unpredictable; instead, it’s about understanding, learning, and being prepared. And that is the heart of true scientific discovery.

The study is published in the Journal of Geophysical Research Solid Earth.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–