July 14, 2013
Stanford scientists eavesdrop on erupting volcano's astonishing seismic sound
Stanford geophysicists listened in on the 2009 eruption of the Redoubt Volcano outside Anchorage, Alaska. By studying and modeling the accelerating earthquakes preceding the volcano's blasts, the scientists hope to better predict the behavior of future volcanic eruptions.
By Thomas Sumner
When Mount Redoubt near Anchorage, Alaska, began spewing ash in 2009, scientists at the Alaska Volcano Observatory and the University of Washington recorded the event using sensors deep inside the volcano. (Photo: Cyrus Read / U.S. Geological Survey)
When volcanoes grumble, scientists listen.
In 2009, Redoubt Volcano outside Anchorage, Alaska, began spewing towering ash plumes more than 12 miles tall. While similar volcanic outbursts are common in Alaska, seismic sensors listening to the volcano's innards recorded something unusual: an accelerating series of earthquakes leading up to each of the volcano's eruptions.
Scientists at the Alaska Volcano Observatory and the University of Washington took their recordings to Stanford University's Eric Dunham, an assistant professor of geophysics, to find out what was happening deep in the volcanic depths of Mount Redoubt. Dunham and his team published their findings July 14 in Nature Geoscience.
"These seismic observations offer a window into the magmatic system and forces acting at several kilometers' depth – arguably the region that's sourcing the magma for these eruptions," said Dunham. "The better we understand the plumbing system beneath the volcano and how magma is being transported, the more accurately we can develop models to predict the timing, duration and explosiveness of eruptions."
When 30 minutes of seismic recordings of an earthquake swarm are sped up to 60 times their original speed, the earthquakes sound a bit like a bag of popcorn in a microwave oven. Each "pop" is a seismic tremor of around magnitude 1.0.
These tremors, while too weak to be felt or heard by someone on the surface, were audible to the sensitive seismic instruments placed along the active volcano's slopes.
Ten hours before each eruption the earthquakes became more and more frequent – like a quickening drumbeat. Just before eruption, sensors were measuring 30 new earthquakes starting every second.
"At the start of this earthquake swarm there were only a few events per minute," said Dunham. "Gradually the earthquakes became more frequent to the point where waves from one event began overlapping with waves from the next event and ultimately merged into a continuous tremor."
This continuous tremor increased in pitch as more and more earthquakes stacked one on top of another. Alaskan seismologists nicknamed these unusually high-pitched tremors, played back at high speeds, "The Screams." But then, seemingly inexplicably, the earthquakes stopped. For 30 seconds the instruments recorded nothing but seismic silence. Finally, the quiet made way for the roar of an eruption. Twenty similar eruptions were measured over Redoubt's two weeks of activity.
"We thought, 'Wow, there's something really interesting happening here –something is driving this system incredibly rapidly,'" said Dunham.
Champagne and molten rock
After studying the seismic recordings, the Stanford team developed a potential explanation. Magma chambers deep inside the volcano are connected to the surface through a narrow passageway called a conduit. A large boulder or a piece of collapsed wall 65 feet across could plug the conduit like a cork in a champagne bottle, building up pressure in the magma below.
The growing pressure below this obstruction would eventually force it upward, grinding and sliding against the conduit walls until it once again became jammed. Each of these scraping motions would cause a small earthquake like the ones measured by the seismic sensors.
"The same thing happens when you run your fingers over a chalkboard – you get stick-slip motion at your fingernail," said Dunham. "That's at a much higher frequency, but it's the same type of thing happening at Redoubt."
As the pressurized magma pushes harder and harder on the obstruction, the quakes become more frequent.
"The fault was being loaded 10 orders of magnitude faster than California's San Andreas Fault," said Dunham. "Instead of getting earthquakes every 200 years, you get 20 per second."
Eventually, instead of sticking and slipping, the obstruction just slides continuously. Dunham believes the obstruction was sliding smoothly up the conduit during the mysterious 30 seconds of seismic silence before each eruption, and a mathematical model of frictional sliding developed by the group backs up this interpretation.
"I don't know if we got lucky and this obstacle just happened to be there," said Dunham. "This particular type of tremor has rarely been reported, but I wouldn't be surprised if it might be seen in other eruptions too."
Using this stick-slip explanation, the team created a computer model of the volcanic earthquakes and forces. This model could be applied to future eruptions to better predict and understand volcanic activity.
"If there are processes leading to a volcanic eruption, you want to detect them early on and try to forecast the explosion," said lead author Ksenia Dmitrieva, a doctoral candidate in geophysics at Stanford. "It's very important to know about volcanoes because they're such a major hazard."
Funding for this project was provided by the National Science Foundation. The research was conducted alongside scientists from the Alaska Volcano Observatory and the University of Washington.
Thomas Sumner is an intern at the Stanford News Service.