Surprise eruptions are Earth’s overlooked threat

Hunga Tonga-Hunga Ha'apai erupted on January 15 in one of the largest volcanic events in decades, and research in recent years indicates that the volcano produces similar violent eruptions about once every millennium. But some of the deadliest and most explosive eruptions in the last two centuries have come from volcanoes with no historical record of eruption. Now, scientists are taking a fresh look at surprise eruptions and seemingly quiet volcanoes.

Several smaller eruptions have happened at Tonga over the last century, providing some clues for what to expect. But forecasting volcanoes that have been dormant for hundreds or thousands of years is especially challenging. Mount Lamington in Papua New Guinea, for example, killed nearly 3,000 people and displaced thousands more when it erupted for the first time in recorded history in 1951.

"If this were to happen right now, in a region that is highly populated, like, for example, Indonesia or the Philippines, where millions of people live around volcanoes, the consequences for life would be really high," said Vanesa Burgos, a PhD candidate at Nanyang Technological University (NTU).

Around the world, over 1,350 potentially active volcanoes simmer beneath the surface. To accurately forecast an eruption, experts need to understand a volcano's past activity. However, records of hundreds of young volcanoes lack detail partly due to little or no monitoring, e.g., local systems to measure earthquake frequency or changes in off-gassing. About 40 percent of active volcanoes do not have a historical record of eruption, making new ones difficult to anticipate.

When Burgos was growing up in Spain, she wanted to study tornadoes. "But you know, tornadoes in Spain are not a thing," she said. Trading one type of disaster for another, she did an assignment on the violent 1991 Pinatubo eruption and was impressed by the difference volcanologists were able to make. Though the volcano had no historical record of eruption prior to that, lives were saved thanks to monitoring systems that were in place.

Now, Burgos and colleagues have created the First Recorded EruptionS in the Holocene (FRESH) database to help identify other areas that could benefit from this sort of proactive monitoring.

Finding areas prone to surprise eruptions

Other databases catalogue information about active volcanoes, such as the Smithsonian Institution's Global Volcanism Program (GVP) Holocene Volcano List, which is a benchmark for volcanologists. But Burgos and her colleagues found that various regions have differing degrees of under-recording, often influenced by historical events such as colonization, submarine volcanism, remoteness, and the level of human development, among other factors.

The researchers at NTU created the FRESH database to narrow the focus and help identify Asia-Pacific regions with higher chances of being surprised by an eruption from volcanoes that are geologically young. In this case, "young" means a volcano that hasn't erupted during last 12,000 years—the geological epoch known as the Holocene.

To do this, the team extracted the first known eruption from each active volcano listed in the GVP's Holocene Volcano List, then created a database of young volcanoes without known Holocene eruptions. Comparing these to analogous volcanoes with previous surprise eruptions helped the researchers estimate the probability of a FRESH eruption occurring at volcanoes with no historical record. The researchers found that some regions particularly prone to under-recording that had higher FRESH eruption rates included Indonesia, Sumatra, the Philippines, and the Mariana Islands. Many of these areas have high populations, and an unexpected eruption of a volcano could be disastrous.

Burgos cautioned that this is preliminary work. "We're not trying to shout 'fire' in the room," she said. "We don't know if these volcanoes are going to erupt anytime soon. We're just saying this is the probability that they might erupt, and this is how many people live around them."

The challenges of accurate volcano forecasting

Burgos said volcanologists logically tend to focus on frequently active volcanoes near highly populated areas. However, she said there's a need for more study of these young, inconspicuous volcanoes. Determining their potential threat requires filling in the gaps in data. "This is a major problem, because the future activity that we want to forecast is mainly based on the past activity that we have seen," Burgos said.

Forecasting eruptions is a complicated task, made more so for these volcanoes. With the help of the FRESH database, which is scheduled to be publicly available this year, Asia-Pacific regions can better understand the probability of their area experiencing an eruption of a FRESH volcano. However, knowing if a volcano might erupt is just the first step.

It's not just a question of if a volcano will erupt, but how big, how deadly, and how far its debris will reach, according to Heather Wright, a research geologist with the Volcano Disaster Assistance Program (VDAP) at the United States Geological Survey. These other considerations will need to be taken into account, especially in terms of preparedness. While the FRESH database wasn't made specifically for policymakers and emergency managers, Burgos said the results could help them prioritize which areas might need deeper study or more consistent monitoring.

John Ewert, a geologist who has been with the VDAP since it was formed in 1986, likes to compare forecasting volcanoes to doctors knowing their patients. "You want to know what the baseline behavior of the system is you're looking at," he said. "Some volcanoes produce earthquakes all the time and do nothing for the most part, but they're active. And others sit there extremely quietly."

Ewert helped update the volcano early warning system and national threat assessment for the U.S. and said long-dormant volcanoes are tough to forecast. Even when they show signs of "reawakening," there's still no real baseline to measure against when they haven't erupted in thousands of years. But Hunga Tonga-Hunga Ha'apai and most other volcanoes do not have monitoring systems installed.

Sensors providing a continuous data stream could establish a minimal baseline, which might help improve forecasting, Ewert said. "You haven't had an annual checkup on this volcano, and so you don't know if it's blood pressure is high or low or if it's got some other factor which is making it more likely for something to happen," he said. "So having that broad sweep of baseline information for potentially active volcanoes would be a key element in terms of monitoring capabilities."

Data isn't just limited to spreadsheets and seismograms though. Wright said there's valuable information in the oral traditions of indigenous communities that surround volcanoes.

"Even the 7,500-year-old climactic eruption of Crater Lake is recorded in the Klamath Indian and Modak Indian oral traditions and legends," she said. "There's so much more to be learned from just enriching our understanding from learning from these long, long-lived stories that have been passed on for so many generations."