Stanford Report Online

Stanford Report, January 10, 2001
The art and science of predicting volcanic eruptions


Volcanic eruptions and earthquakes are two of the most destructive forces on Earth.

But while scientists can do little more than guess when an earthquake will strike, tremendous strides have been made in forecasting deadly volcanic explosions.

A dramatic example occurred on Dec. 18, when volcanologists accurately predicted the eruption of Popocatepetl near Mexico City just hours before it underwent its most violent explosion in centuries. The early warning gave Mexican officials enough time to evacuate some 40,000 people living in the shadow of the volcano.

Hot lava explodes from Hawaii's Pu`u `O`o Volcano – one of nearly 60 volcanoes that erupt worldwide every year. Courtesy: J.D. Griggs, US Geological Survey, Hawaiian Volcano Observatory.

Coincidentally on Dec. 15, three days before Popocatepetl blew its top, a special session on the future of volcano prediction was held at the fall meeting of the American Geophysical Union (AGU) in San Francisco. Panelists from the United States, Iceland and Germany discussed advances in technology and communication that could revolutionize the entire field of volcanology in the coming decade.

"It's not so much that we expect one technical breakthrough," observed Stanford geophysicist Paul Segall, who co-moderated the half-day AGU session. "There's a confluence of a lot of things going on that should improve our ability to make predictions in the next 10 years," he said.

Segall pointed out that earthquakes often foreshadow major volcanic events.

"You can't move lots of magma through the Earth's crust without earthquakes," he said, noting that volcanologists have developed specialized seismic arrays for imaging volcanic systems.

To predict eruptions, volcanologists are using innovative technologies that measure subtle changes in the shape, temperature and gas chemistry of active volcanic vents such as Pu`u `O`o on Hawaii's Big Island.
Courtesy: C. Heliker, US Geological Survey, Hawaiian Volcano Observatory.

Segall and Stanford colleague Howard A. Zebker, an associate professor of electrical engineering and geophysics, described the increasingly popular use of spaceborne satellites to monitor volcanic activity on Earth.

Tiny movements on the surface of a volcano often indicate the buildup of magma below. Segall and Zebker speculated that, in the next decade, the Earth may be orbited by an array of specially equipped radar satellites capable of detecting millimeter-sized changes in the Earth's crust.

"Improved satellite coverage will make it possible to collect these data over all of Earth's 600 potentially active volcanoes weekly or even daily," they predicted, allowing researchers to forecast volcanic events to a much greater degree than is currently possible.

"Volcanoes tend to be dangerous," Zebker told reporters following the AGU session, noting that rough terrain and unexpected eruptions have cost many scientists their lives. "We're better off doing things remotely."

According to the U.S. Geological Survey (USGS), approximately 30,000 people have been killed by volcanoes in the last 20 years, including some 25,000 victims of the 1985 eruption of Nevado del Ruiz in Colombia.

Arizona State University geologist Jonathan Fink, who co-moderated the panel, pointed out that several big questions face volcanologists trying to predict an eruption: "Is magma inside the volcano rising? How much? How fast? How much gas is being emitted? Will magma continue to rise or will it stop?"

Fink, who earned his doctorate at Stanford in 1979, ominously noted that, in the next decade, at least four volcanoes have the potential for major eruptions with many fatalities: Rainier in Washington State, Vesuvius in Italy, Taal in the Philippines and Popocatepetl.

Panelist Bernard Chouet of the USGS described some of the latest advances in volcano seismology, including portable broadband seismic instrumentation, high-resolution tomography to create images of underground volcanic structure and other tools to measure the acoustic properties of magma and hydrothermal fluids.

Several researchers discussed the use of global positioning systems (GPS) to detect bulging of the ground, as well as geochemical devices to monitor sudden changes in atmospheric gas chemistry that often precede eruptions.

Andrew J.L. Harris and his colleagues from the University of Hawaii described their pioneering thermal monitoring system that records second-by-second temperature changes at Hawaii's Pu'u O'o volcanic vent, then instantly transmits the data to a computer via satellite.

"We advocate that such remote systems should be installed on other volcanoes by 2010 in order to better monitor ongoing eruptions," Harris said.

Scientists from NASA's Jet Propulsion Laboratory explained the use of digital animation to illustrate potential volcanic hazards, and researchers from Los Alamos National Laboratory and the Pacific Disaster Center discussed ways to improve volcanic risk assessment.

"Much work remains to be done toward a synthesis of seismological, geochemical and petrological observations into an integrated model of volcanic behavior," Chouet concluded.

Segall agreed and expressed optimism about volcanology's future.

"Unlike earthquake prediction," he said, "we can actually do something about volcanoes."