Machine learning improves volcano eruption warnings
Analysis based on 7 articles · First reported Aug 30, 2026 · Last updated Sep 01, 2026
The study could influence investment in volcano monitoring technology and early warning systems, potentially benefiting companies in the disaster tech sector. It may also affect tourism and infrastructure planning in volcanic regions, as improved warnings could reduce economic losses from eruptions.
Researchers from the University of Canterbury, including Alberto Ardid, David Dempsey, Shane Cronin, and Tony McDonald, published a study in The Conversation proposing the use of machine learning to improve early warning systems for sudden volcanic eruptions. The study re-analyzed the performance of a machine-learning eruption forecaster at five volcanoes in New Zealand, Japan, Chile, and Russia, using years of seismic data to estimate eruption probability within a rolling 48-hour window. At Whakaari/White Island, the forecaster anticipated four of five eruptions, with a tradeoff of about 15 days per year of false warnings. Using a cost-loss model, the researchers found that precautionary action could reduce preventable losses by 30-90%. The study argues that accepting more false alarms may be economically rational and safer, drawing parallels to tsunami warning systems. It emphasizes that automated warnings would complement, not replace, human volcanologists and aligns with the United Nations' Early Warnings for All initiative. The research highlights the need to balance economic costs, such as closing New Zealand — Mount Ruapehu during ski season, against the potential to save lives.
Set up alerts, explore entity relationships, search across thousands of events, and build custom intelligence feeds.
Open Dashboard