Title: Operational Eruption Forecasting: A Global Perspective
Historically, volcano eruption forecasting has been based primarily on seismology. Traditional seismological forecasting techniques, which involve tracking the temporal evolution of seismic events with various source processes (brittle failure earthquakes, long period earthquakes, and tremor) in context of other data streams, has been successful for many high-impact eruptions. However only 22% of historical eruptions in the US were preceded by volcanic swarms and only 33% of eruptions at monitored volcanoes in Alaska since 1990 were successfully forecast. In this talk I will describe common methods of eruption forecasting based on seismicity and give examples of when these methods succeed and when they fail, based on the practical experience of the Alaska Volcano Observatory (AVO) and the USGS Volcano Disaster Assistance Program (VDAP). AVO and VDAP are both engaged in multiple eruptions annually with a wide variety of characteristics. Thus, the combined work of these organizations provides a realistic global snapshot of how new seismology research can potentially meet the operational needs of volcano observatories. In context of these lessons, I will present four examples of new efforts between VDAP and its partners to improve eruption forecasting and monitoring: 1. Monitoring remote volcanoes (Hunga Tonga-Hunga Ha’apai, Kingdom of Tonga, 2022), 2. Evaluating earthquake swarms (database of all volcanoes in the US), 3. Insight from tracking earthquake families (Agung volcano, Indonesia, 2018), 4. Practical use of infrasound and alarms (Fuego, Guatemala, 2018).