Eruption Disruption: Exploring New Volcano Monitoring Technology
By Sierra McConnell
Researchers have made incredible strides in monitoring volcanic activity over the last decade, making these natural phenomena easier to spot and predict. In turn, they are reported on more often, giving the illusion that volcanic eruptions are on the rise when they have remained about the same for over 20 years.
At the time of this writing, 64 volcanoes have erupted in 2024. To put that into perspective, there have been between 67 and 83 volcanoes with reported activity each year since 2000.
Monitoring on the Ground and in the Air
All volcanic monitoring focuses on detecting deviations in activity, ranging from simple observations like lava flow to advanced techniques like drone monitoring, GPS network signaling, and machine learning for eruption prediction.
One new tool is the United States Geological Survey (USGS) Spider, a package of several monitoring instruments that can be safely deployed from a helicopter. This device reduces the risk to scientists and enables rapid data collection in remote or hazardous areas. Designed to monitor seismic activity, ground deformation, and gas emissions, the Spider, which gets its name from its leggy appearance, provides comprehensive data in real time. Its ability to integrate various monitoring tools into a single unit lowers power requirements, minimizes environmental impact, and has vastly improved the safety and efficiency of volcano monitoring efforts across the globe.
Another group of researchers at Cornell University discovered that, by analyzing the fluid found within volcanic crystals, they can determine magma’s location before, during, and after an eruption with an accuracy of up to 100 meters. This technique allows for rapid and precise data collection, significantly enhancing the speed and accuracy of eruption forecasts. Such advancements enable scientists to provide near real-time assessments, which are crucial for timely evacuation and risk management to nearby populations.
Long-range drones equipped with miniaturized gas sensors, spectrometers, and sampling devices have also transformed data collection in hazardous volcanic environments. These unmanned aerial vehicles can now sample gases directly from volcanic plumes, significantly improving researcher safety and data quality. Cambridge University’s Aerial-Based Observations of Volcanic Emissions (ABOVE) project exemplifies this advancement, having successfully deployed drones to study the highly active Manam volcano in Papua New Guinea.
"Machine learning algorithms significantly enhance volcano monitoring and eruption prediction."
Data Integration and Machine Learning
A study published in Frontiers in Earth Science demonstrated the potential of machine learning algorithms to significantly enhance volcano monitoring and eruption prediction. Researchers developed a novel approach that analyzes four key seismic features: energy, softened Shannon entropy, kurtosis, and frequency index. By applying this method to data from various volcanoes, the team created a probabilistic tool for real-time monitoring that can provide early warnings from hours to days in advance. This innovative technique represents a major step forward in volcanic hazard assessment and has the potential to greatly improve disaster preparedness and risk mitigation efforts.
The system works by organizing the seismic data into a temporal matrix, which captures the evolution of volcanic activity states over time. This matrix is then processed by a neural network, a sophisticated machine learning model capable of recognizing complex patterns that traditional methods might overlook. The neural network learns to identify precursory signals of volcanic activity by analyzing the relationships between the four seismic features. Once trained, the network can process new data in real-time, generating probabilistic forecasts of imminent eruptions. This approach not only enhances the accuracy of eruption predictions, but also provides volcanologists with a powerful tool for continuous, automated monitoring of volcanic activity.
Future Directions
With advancements in AI-driven forecasting and durable equipment enhancing eruption predictions, the future of volcanic monitoring is bright. As technologies evolve, they offer unprecedented insights into volcanic behavior, improving the ability to mitigate risks and protect communities living near these sometimes-terrifying natural wonders.
Sierra McConnell is a Thermo Fisher Scientific staff writer.
