Science News logo

Science Service logo

When Mountains Fizz
What makes some volcanic eruptions so explosive?

Sid Perkins

On May 18, 1980, Mount St. Helens blew its top. After the northern face of the mountain slumped away in the most massive landslide ever recorded, molten rock gushed up from reservoirs several kilometers below the exploding peak. Within minutes, a plume of volcanic ash surged more than 20 km into the sky. Immense flows of mud and ash covered 62 square kilometers, devastated forests, rivers, and lakes, and killed 57 people.


TEMPERAMENTAL PEAKS. The concentrations of dissolved gases in a volcano's molten rock largely determine whether an eruption is explosive, like the Mount St. Helens eruption of May 18, 1980 (above, in close-up), or effusive, like the eruption at Hawaii's Kilauea, which has been going on for more than 2 decades (below).

D. Swanson, USGS

Contrast that event with the latest activity of Hawaii's Kilauea volcano, an event that began early in 1983 and is still under way. Although the mountain occasionally spews fountains of lava hundreds of meters high or burps small clouds of ash, the molten rock it exudes usually just flows down the slopes. Sometimes, the magma races along in torrents, but at other times the flow is so slow that scientists can safely walk up to the searing trickle and pry away fresh samples of congealing rock.


C. Heliker, USGS

Two volcanoes. Two eruption styles. Geologists have long sought to uncover what underlies such variation, which sometimes gives a volcano different temperaments at different times. Scientists have searched for answers in the physical structure and chemical composition of lava and ash, but some geophysicists have begun looking under other sorts of scientific stones.

While some of these researchers use gooey concoctions and elaborate test equipment to replicate and measure magma flow in the laboratory, others use sophisticated computer models to simulate volcanic eruptions. And the more these researchers delve into these pursuits, the more convinced they become that Mount St. Helens, Kilauea, and the rest of the world's volcanoes have a lot in common with soda pop.

Thar she blows!

At the great pressures deep within Earth, molten rock can hold high concentrations of dissolved gases. As gas-charged magma makes its way to Earth's surface, where pressures are much lower, the gases—water vapor, carbon dioxide, sulfur dioxide, and others—begin to form bubbles within the material. As the molten rock continues its ascent, those bubbles grow and the magma expands and becomes less dense, says Michael Manga, a volcanologist at the University of California, Berkeley.

The sudden landslide that preceded the eruption on Mount St. Helens in May 1980 was like popping the top on a soda can that's been shaken. The abrupt relief of pressure on the gas-charged magma beneath the peak created bubbles in the molten rock, and the liquid around the bubbles became extremely viscous. In such a state, the magma acts like taffy—when stretched slowly, it yields, but if it's pulled quickly, it breaks apart into ragged fragments.

When the rapidly expanding bubbles carried the viscous fluid up through the throat of Mount St. Helens, the magma blew apart as it breached the surface. That eruption ultimately ejected more than 1 cubic kilometer of volcanic ash, U.S. Geological Survey scientists estimate. Geologists define ash as volcanic particles less than 4 millimeters in diameter.


STRETCHED THIN. Rock in samples of obsidian collected at California's Big Glass Mountain can take different forms: fragmented and reconsolidated (top), slightly elongated (center), or stretched into thin layers (bottom, close-up). This progressive elongation reflects the forces that shape the taffylike magma inside a volcano.

H. Gonnermann

Just a few years ago, most scientists thought that the fragmentation of magma occurred only in explosive eruptions, says Manga. In quiet, effusive eruptions, they assumed, the molten rock either didn't contain much dissolved gas or the bubbles found a way to escape the magma without blowing it apart. However, Manga told a meeting of the American Geophysical Union (AGU) in Montreal last May that his group's analyses of volcanic material collected at California's Big Glass Mountain suggest that magma fragmentation can occur deep within volcanoes without triggering an explosive eruption.

Think of fluid moving in a pipe, says Manga. Friction makes the fluid flow more slowly near the walls of a pipe than in the middle. Physicists call this phenomenon the boundary effect. Similarly, in molten rock flowing through the throat of a volcano, the shear forces due to friction, which tend to rip apart the material, are greatest near the rocky walls of the conduit, says Manga. Under high pressure deep inside a volcano, he and his colleagues reason, magma is repeatedly torn into fragments by such forces and then squeezed together again.

Indeed, Manga and his colleagues found evidence of the repeated fragmentation and consolidation of magma when they examined samples of banded obsidian, a type of volcanic glass, formed during an ancient eruption of Big Glass Mountain. Some samples of the obsidian included many lumps of fragmented rock embedded in the glass. In other samples, the fragments were elongated as if they'd been stretched. In a third set of samples, the fragments had been stretched so much that they were transformed into long, thin smudges, stacked together so that in cross section they showed up as narrow bands.

Besides supporting the idea that fragmentation can occur during nonexplosive volcanic eruptions, these observations offer a scenario for the formation of banded obsidian, Manga notes.

In some effusive volcanoes , networks of small fissures formed by magma fragmentation in the throat of a volcano serve as relief valves for gases that otherwise would propel an explosive eruption, Manga and his colleagues speculate. The ash-filled cracks found in the glassy walls of a previously active conduit of Iceland's Torfajökull volcano bolster this notion, says Harry Pinkerton of Lancaster University in England. The layers of ash and other mineral fragments clogging those crevices suggest that the material had been carried by episodic bursts of hot gases through a system of fractures.

The size and some other characteristics of these cracks, which appear a few dozen meters below the volcano's vent, suggest that the fissures may have stayed open for several seconds during each cycle of fragmentation. They would thereby relieve pressure and prevent an explosive eruption, Pinkerton and his colleagues reported in 2003.

Got gum?

Because it's dangerous and difficult to investigate volcanic eruptions in progress, scientists are devising a host of experiments to simulate the flow of magma in the laboratory.

Manga and his Berkeley colleague Atsuko Namiki, for example, are working to solve the riddle of why volcanoes with the same magma chemical composition and dissolved gas content erupt explosively in some cases and effusively in others. The researchers are conducting experiments with a viscous material called xanthan gum, a thickener found in such food items as yogurt, pudding, and salad dressing. It's convenient in the lab because it's nontoxic, inexpensive, and available in bulk. It's also easy to adjust the material's viscosity by controlling its water content, Manga notes.

First, Manga and Namiki used an eggbeater to whip up various slurries of xanthan gum. Then, they placed the bubble-rich blends in clear tubes, sealed the ends, and applied pressure. High-speed videotapes revealed that when the pressure was suddenly released, the size of the bubbles didn't influence whether the subsequent eruption was explosive or effusive.

However, the researchers found that the potential energy stored in the bubbles—equivalent to the total volume of a mixture's bubbles multiplied by the extra pressure placed on the mixture in the chamber—affected the expansion rate of the suddenly depressurized fluid.

In low-viscosity mixtures or at slow expansion rates, the bubbles escaped the xanthan gum preparation before stresses in the fluid fractured the faux magma—resulting in a Kilauea-style, effusive eruption. Faster expansion rates promoted fragmentation of the fluid and a violent eruption, à la Mount St. Helens, says Manga.

In other experiments, Jacopo Taddeucci of the National Institute of Geophysics and Volcanology in Rome and his colleagues used the toy called Silly Putty as a magma substitute. After saturating the material with the inert gas argon at high pressure in a sealed chamber—similar to adding gas to a can of soda—they suddenly released the pressure.

When the bubbles formed and expanded rapidly, the putty fractured in an explosive nanoeruption. Once the gas had escaped, however, the fractures in the putty began to heal. The same fracturing, gas venting, and healing processes probably occur during volcanic eruptions and in part determine whether the eruption is explosive or effusive, says Taddeucci.

Lab tests aren't restricted to artificial magma. Taddeucci and his colleagues have also used reheated samples of volcanic material that had been flung from Italy's Mount Etna during its explosive eruptions in the summer of 2001. Among the samples the researchers collected were rocks 40 centimeters across that had landed 450 m from the volcano's vent.

Back in the lab, the researchers drilled small, cylindrical samples from the rocks and measured their porosity using inert helium gas. This technique doesn't affect the samples' chemical composition or crystallography, says Taddeucci. Then, he and his colleagues subjected the samples to conditions of 850°C and up to 250 times normal atmospheric pressure (atm). Even though the temperature in these tests was about 150°C lower than the magma's eruption temperature, the test results should be relevant to Mount Etna, Taddeucci notes.

When the pressure was suddenly released from rock samples that had been pressurized to 50 atm or more, the material fractured. This suggests that pressure in the magma during the Mount Etna eruption couldn't have been more than 50 atm because otherwise the rocks wouldn't have remained intact. However, Taddeucci notes, pressure in the magma had to be at least 25 atm to fling the samples so far from the vent. Therefore, inside the erupting volcano, the pressure must have been between 25 atm and 50 atm, the researchers concluded in the Aug. 10, 2004 Journal of Geophysical Research (Solid Earth).

Data, please

Simulations of volcanic eruptions with an analog such as xanthan gum and tests on actual volcanic materials both supply data useful to researchers who develop computer models of eruptions.

Most models are simplistic, says Jeffrey B. Johnson, a volcanologist at the University of New Hampshire in Durham. It's easy to represent some factors, such as the pressure in magma at various depths in a volcano. However, most models ignore details of fiendishly complicated parameters, such as changes in the magma's viscosity as a function of temperature.

Nevertheless, says Johnson, many models do a good job of simulating certain types of eruptions. In a steady, explosive eruption, for example, the fragmentation front—the narrow zone in which bubbles blow the rapidly ascending molten rock into ash particles—remains at a particular depth in the throat of a volcano.

Slightly more complicated is the episodic eruption, in which the volcano alternates between extended periods of quiescence and bursts of explosive activity. Geologists wonder how the pressure in the magma builds and what's happening in the volcano's magma tube during that quiescent interval.

Alison C. Rust, a volcanologist at the University of British Columbia in Vancouver, has developed a model that suggests answers to those questions.

In an episodic eruption, she says, experiments indicate that magma contains only a moderate supply of dissolved gases, including water vapor. In her model, once the eruption begins, the fragmentation front travels downward into the throat of the volcano much faster than the magma rises. As a consequence, the eruption—figuratively and literally—runs out of steam, exhausting the supply of water vapor that's driving the fragmentation. Until enough bubbles again accumulate, the eruption is placed on hold.

By looking for similarities among different models of volcanoes, Dork L. Sahagian of Lehigh University in Bethlehem, Pa., identified what geophysicists generally agree are the important factors influencing eruptions. Those parameters also form a volcanologist's wish list of where future experimentation and analysis should be focused, he proposed at the AGU meeting in Montreal.

First, modelers need to better understand how magma flows during bubble growth and how it behaves immediately before it's blown to fragments in an eruption. Although researchers have collected data on how viscous molten rock flows when it's stretched at slow rates, little if any data have been reported on how magma flows at high strain rates, Sahagian notes.

Second, scientists must explore volcanoes' internal plumbing. Volcanoes are often modeled as long tubes that lead from the underground magma reservoirs to a single crater atop a symmetrical cone rather than as a convoluted conduit that leads to a branching network of mountaintop vents. Monitoring the pattern of seismic activity around volcanoes could illuminate details of a the mountains' inner structures, which then could be incorporated into more-realistic models, Sahagian suggests.

Finally, researchers should investigate the origins of the gas bubbles that ultimately drive volcanic eruptions, he says.

Ironically, even though the amount of dissolved gases in magma is one of the most important factors in determining whether any particular eruption will be a Mount St. Helens–size boom or a Kilauean bust, scientists know little about how bubbles of gas form inside molten rock, Sahagian notes.



Gonnermann, H.M., and M. Manga. 2003. Explosive volcanism may not be an inevitable consequence of magma fragmentation. Nature 426(Nov. 27):432–435. Abstract.

Johnson, J.B., A. Proussevitch, and D. Sahagian. 2004. Modeling transient eruptive processes. Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.

Manga, M., H.M. Gonnermann, and A. Namiki. 2004. Why do volcanoes (only sometimes) erupt explosively? Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.

Namiki, A., and M. Manga. 2004. Importance of preexisting bubbles for volcano explosivity. Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.

Rust, A.C., K.V. Cashman, and K.M. Wearn. 2004. Degassing and fragmentation in sustained vs. episodic eruptions: Evidence from pyroclastic obsidian. Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.

Sahagian, D. 2004. Volcanic eruption mechanisms: Insights from intercomparison of models of conduit processes. Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.

Taddeucci, J., et al. 2004. Experimental and analytical modeling of basaltic ash explosions at Mount Etna, Italy, 2001. Journal of Geophysical Research (Solid Earth) 109(Aug. 10):B08203. Abstract.

Taddeucci, J., et al. 2004. Vesiculation, flow, and fragmentation of viscoelastic magma: Inferences from an American toy. Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.

Tuffen, H., D.B. Dingwell, and H. Pinkerton. 2003. Repeated fracture and welding of viscoelastic magma rising in volcanic conduits. European Geophysical Society and American Geophysical Union meeting. April 6–11. Nice, France. Abstract.

Further Readings:

Spieler, O., B. Scheu, and D.B. Dingwell. 2004. From magma to pyroclasts: Experimental approaches to the understanding of explosive eruptions. Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.

Tuffen, H., D.B. Dingwell, and S. Sturton. 2004. Shear fracture and healing of silicic magma during conduit flow: Implications for seismicity and degassing. Canadian Geophysical Union and American Geophysical Union meeting. May 17–21. Montreal. Abstract.


Jeffrey B. Johnson
University of New Hampshire
Climate Change Research Center
Morse Hall
Durham, NH 03824

Michael Manga
University of California, Berkeley
Department of Earth and Planetary Science
Berkeley, CA 94720

Harry Pinkerton
Room A 513
Department of Environmental Science
Institute of Environmental and Natural Sciences
Lancaster University
Lancaster LA1 4YQ
United Kingdom

Alison C. Rust
University of British Columbia
Department of Earth and Ocean Sciences
6339 Stores Road
Vancouver, BC V6T 1Z4

Dork Sahagian
Department of Earth and Environmental Sciences
Lehigh University
31 Williams Drive
Bethlehem, PA 18105
Web site:

Jacopo Taddeucci
Department of Seismology and Tectonophysics
National Institute of Geophysics and Volcanology(INGV)
Via de Vigna Murata 605
Rome 00143

From Science News, Volume 167, No. 5, January 29, 2005, p. 74.