ASU Research E-Magazine: Predicting The Blast

The immense natural power of erupting volcanoes has obliterated entire islands and buried cities. The ability to predict volcanic explosions with any degree of accuracy could save many lives. ASU scientists are among those working to develop a reliable forecasting technology.

Sometimes the best thing a professor can do for an inquiring young scholar is to leave him alone.

That’s exactly what happened to Arizona State University graduate student Tobias Fischer last year. On Jan. 14, 1993, a tragic eruption at the Galeras volcano in South America took the lives of six scientists. The eruption severely injured Fischer’s mentor, ASU geologist Stanley Williams.

Fischer’s largely unsupervised research at Galeras with two fellow ASU students produced a groundbreaking article in the March 10, 1995 issue of Nature, one of the world’s top scientific journals.

He and his co-authors did what no volcanologist had done before: they compared the earthquake data from a volcano to its gas data. They found that monitoring volcanic-gas emissions in connection with seismic events may help forecast future eruptions.

“People who study the seismology of volcanoes don’t talk very much to people who study the gases. They don’t speak to the people who do the gravity studies. And they don’t speak to the geologists who just look at the rocks,” Williams says.

“We get too superspecialized, really. Maybe Toby was naive enough to not know that he shouldn’t cross that boundary. He compared the seismic events beneath the volcano with the gas measurements and he came up with a correlation.”

A native of Germany, Fischer received his bachelor’s degree from Albrecht-Ludwigs University in Freiburg. He spent much of the summer of 1992 in South America, studying volcanoes in Ecuador and Colombia, including Galeras.

After the January 1993 accident, Fischer spent some of his days collecting water and gas samples at Galeras and other volcanoes in the area. Other days he would remain at the Pasto Volcanological Observatory, analyzing Galeras seismic data. If the seismic data indicated the volcano was stable, he could go into the field to collect samples. He also measured sulfur dioxide gas emissions with an instrument called a correlation spectrometer (COSPEC).

The relatively mild earthquakes that occur beneath a volcano go by the term “long-term events” in volcanology lingo. The special name sets them apart from the more familiar “tectonic” earthquakes that result from movements along faults.

Fischer knew that, in theory, scientists had linked long-period events with gas movements. He did not know that no one had ever directly compared data from the two phenomena. He probably wouldn’t have compared them, either, under different circumstances.

“I was not able to collect as much gas data as I wanted to because of the accident,” Fischer says. “I wanted to go three times a week, but it turned out it was much too dangerous. So I was looking for something else that I could put into my master’s thesis.”

The Pasto observatory had seismic data leading up to the Jan. 14 eruption, but no gas data. Fischer began collecting gas data at Galeras in February and March. The volcano exploded again on March 23. Rising at dawn the next day, Fischer took the COSPEC instrument into the volcano to get post-eruption gas data as well.

He found that gas emissions had dropped rapidly before the March 23 eruption. Galeras then spewed about 1,000 tons of sulfur dioxide the day after the eruption. The figure dropped to 500 tons by the following day. The energy released by earthquakes first increased, then decreased before the eruption.

Gas releases fell after the eruption because the volcano’s cracks were open. The built-up pressure bled off quickly. Seismic and gas activity had dropped before the eruption as well because the volcano sealed up. “Pressure was just sitting behind a membrane, a frontier, some realm in the edifice of the volcano,” Williams explains.

The volcano seals itself off again after an eruption, starting a new cycle. “There’s no release vent anymore,” he says. “There are no cracks open. The pressure builds and builds, and then finally when you reach some critical pressure level, it exceeds the strength of the sealing.

“All the cracks open, and now they open catastrophically. They all open widely at one moment. The gases escape and there’s a big explosion.”

ASU doctoral students Meghan Morrissey and Marta Calvache also were at Galeras for the 1993 workshop. Morrissey contributed her fluid-dynamics expertise to the Nature paper. Calvache provided the geologic history of Galeras. Also listed as co-authors were Williams, Diego Gomez, and Roberto Torres of the Pasto observatory, and John Stix of the University of Montreal.

Morrissey is writing her doctoral dissertation on the relationship between long-period seismic events and pressure changes in subsurface fluids at Alaska’s Redoubt volcano. A swarm of about 4,000 long-period events preceded Redoubt’s eruptions of late 1989 and early 1990. They produced the same seismic signal at regular intervals.

“It’s a cavity resonating in the Earth, a crack,” Morrissey says. “Every so often the fluid conditions in that crack produce seismic signals.”

The crack at Redoubt is buried nearly a mile below the surface, near the volcano’s magma chamber. “We’re pretty sure that the fluids going through there are from the magma itself,” Morrissey says.

There are a series of cracks below both volcanoes, but at Redoubt, one crack seems to be resonating. At Galeras there are many. “It’s very difficult to model the actual fluid dynamics in each one of those cracks to see the true source of the long-period signal,” she adds.

Morrissey’s research adviser is geologist Susan Kieffer, now with the University of British Columbia. Improving volcano forecasting also is their goal.

“What we saw at Galeras may not be the case at Redoubt or other volcanoes, but it’s something that we have to start considering and interpreting,” Morrissey says.

The seismic data and the spotty gas data from other eruptions do seem to match the Galeras pattern, however. Williams cites eruptions in Costa Rica since 1968, Papua New Guinea in 1979, and Alaska’s Redoubt in March of 1990.

“We’ve now got half a dozen examples around the world of an explosion at different volcanoes that seem to be consistent with our idea,” he says. Although he was excited by the original thinking the students displayed in their Nature paper, volcanologists have many mountains still to climb.

“Finding an example or two doesn’t prove that a model is anything but a good idea. It hardly demonstrates that it has a general significance. It’s a provocative paper. It’s not ideal by any stretch,” Williams says.

Williams and his students continue to pursue other avenues of research, just in case. Marta Calvache, for example, is finishing her doctoral dissertation on the geologic history of Galeras. Volcanoes, like people, are better understood in the context of their pasts.

She has found that Galeras is not one volcano but a complex of at least five progressively older volcanoes. Several times Galeras has built a big volcanic cone, which then collapsed, forming a big basin. The latest in the series, which Calvache named “Urcunina” (fire mountain), is about 5,000 years old. The oldest could be up to 2 million years old.

“To say it is a certain age is very difficult because the activity of volcanoes in this part of Colombia and this part of South America has been going on for millions and millions of years,” she says.

A native Colombian and a resident of Pasto, Calvache knows firsthand the fears and horrors an active volcano can produce. She was on Galeras the day of the fatal eruption, and risked her own life to help Williams to safety.

Galeras looms directly over Pasto, a city of 300,000. Calvache compares its residents to those of earthquake-threatened Los Angeles.

“They know they will have an earthquake, but sometimes people say, ‘I don’t care. I am living here. I have my house. I have everything here.’ It’s very difficult to leave.

“As soon as they have an earthquake, the next morning probably 10 percent of the people want to leave. But after a month, two months, people forget everything. It is the same in Pasto.”—Steve Koppes