"One of the central problems of volcanoes is: Why do they erupt and why do they alternate between relatively benign effusive eruptions and destructive explosive eruptions?" said Michael Manga, UC Berkeley associate professor of earth and planetary science. "This hypothesis helps us understand why it happens and how it happens."
The conventional explanation holds that in explosive eruptions only, rising magma breaks or fragments as it approaches the surface, releasing bubbles that blow the magma out like champagne from an uncorked bottle.
Manga and graduate student Helge M. Gonnermann are challenging this explanation, proposing instead that fragmentation occurs in most if not all volcanic eruptions, though non-explosively. They detail their hypothesis in the Nov. 27 issue of the journal Nature.
In an explosive eruption, the magma rises fast, allowing a build-up of gas pressure within the gas bubbles that leads to rapid bubble growth, abrupt fragmentation, and explosive release of gas pressure.
In effusive eruptions, they argue, continuous, repeated fragmentation during the magma's rise to the surface would break gas bubbles and allow the release of significant amounts of gas before the magma reaches the surface, leaving little for an explosion.
In other words, volcanoes explode only when the release of gas by continuous and repeated non-explosive fragmentation during the magma's rise to the surface cannot keep up with the growth of bubbles.
"This non-explosive fragmentation should decrease the chance of an explosive eruption," Manga said. "In some cases, this should let the gas out of the volcano and leave nothing to drive an explosive eruption. The detection of non-explosive fragmentation could therefore have important implications for hazard assessment of volcanic eruptions."
"In the past, people have thought that the defining characteristic of explosive eruptions is fragmentation," Gonnermann said. "While that's true, we're saying fragmentation also plays an important role in non-explosive eruptions. We've shown that it is quite feasible that you will get non-explosive fragmentation under a wide range of observed eruptive conditions."
The standard explanation for explosive volcanism is that as liquid magma rises from a magma chamber at a depth of several miles, dissolved gases - mostly water - form bubbles. These bubbles grow in size, because the pressure in the surrounding magma decreases as it rises toward the surface. As bubbles grow, the overall magma volume increases, causing it to rise even faster.
At the same time, the loss of dissolved gases makes it more difficult for the magma to flow, thereby increasing its susceptibility to break in a brittle manner. The process accelerates until the liquid magma actually breaks or fragments.
"If the magma rises fast enough, it can break up even though it's a liquid," Manga said. "Silly Putty is the same way - it deforms smoothly if you pull slowly, but it breaks if you pull quickly."
Fragmentation, probably at depths of thousands of meters, suddenly releases the large volume of gas contained in bubbles, allowing it to blast fragments out the top, much the way carbon dioxide bubbles push soda out of a bottle that has been shaken before opening.
Scientists have attributed the explosive eruptions of volcanoes such as Mount St. Helens in 1980 and Mount Pinatubo in the Philippines in 1991 to magma fragmentation.
On the other hand, volcanoes which erupt explosively also frequently erupt magmas of similar composition by slowly oozing lava, as was the case after the 1980 eruption of Mount St. Helens.
Gonnermann and Manga showed through calculations that, as a consequence of the flow behavior of the ascending magma, effusively erupting magma should also fragment as it rises.
In fact, Manga said, magma may fragment many times during its rise, releasing gas non-explosively each time and then coming together again as a viscous liquid to continue its rise. The fragmentation at depth is caused by strong shear forces near the walls of the conduit through which the magma rises, while fragmentation that produces an explosive eruption is caused by sudden rupture of bubbles throughout the magma, essentially blowing it apart.
"Our model predicts that shear-induced fragmentation should occur even for slowly ascending magmas," Gonnermann said. "Since the magma contains bubbles, bubble walls will get broken during shear-induced fragmentation. As a consequence, gas - mostly water vapor but also carbon dioxide and sulfur dioxide - can escape more efficiently than it could without shear-induced fragmentation."
The theory helps explain the fact that the water and bubble content of lava from an effusive eruption is considerably less than it should be. Magma deep underground probably has a water content of about 6 percent, while that of the effusively-erupted, glassy black lava rock called obsidian is about 0.2 percent, Manga said.
According to the new hypothesis, magma emerging as such a lava flow would have lost a lot of its water during its rise to the surface. As the water formed gas bubbles, repeated non-explosive fragmentation ruptured bubble walls and allowed the gas to escape easily through the passageways formed by the broken bubbles. As a consequence, the erupting lava has a relatively low content of gas bubbles compared to pumice formed during an explosive eruption.
Pumice can be as much as 98 percent gas bubbles by volume, Gonnermann said. If little water is lost during magma rise, the water eventually fizzes out as pumice in an explosion through the volcano's crater. This foamy volcanic rock constitutes many tephra deposits formed by explosive volcanic eruptions.
The formation of a liquid again after fragmentation, called reannealling, also could explain the banded or layered look of much lava rock, such as obsidian.
"Most obsidian has gray and black bands, but there is no good explanation why," Manga said. The gray bands are often the same composition as the black bands, they just contain more microscopic crystals or bubbles.
Based on their model, the researchers suggest that magma pieces of various sizes produced during fragmentation would be carried along in the magma stream and squeezed and stretched into layers, much as intertwined pieces of colored Silly Putty form thin layers of different colors when stretched.
The idea that fragmentation might occur in all rising magma came to Gonnermann after he began looking at banded obsidian from lava flows, in particular the flow at northern California's Big Glass Mountain, a volcanic dome containing about 1 cubic kilometer of rhyolite, obsidian and pumice from an effusive eruption.
In their paper, the researchers compare three different slices through obsidian from Big Glass Mountain, representing a gradation from rock samples composed of many pieces of fragmented rock to an almost smooth-banded structure with no fragments visible to the naked eye. Under a microscope, however, tiny crystals as small as 10 microns - one-tenth the diameter of a human hair -are visible in the black and gray bands.
Gonnermann and Manga now are trying to find further evidence in obsidian of non-explosive fragmentation. They also are trying to understand how to distinguish "flow" banding that was formed through a fragmentation process from banding produced solely through viscous flow of liquid magma. This has important implications for understanding how magmas of different compositions can mix together.
The viscosity of the magma also determines whether the rising column fragments or not, they said. A much less viscous magma than those considered by Gonnermann and Manga allows bubbles to rise through it to the surface. At the same time, it requires a much higher flow rate to break, so volcanoes with less viscous magma, such as Hawaiian basalt, usually erupt non-explosively.
The research is funded by the National Science Foundation and the Sloan Foundation.