How do You Know if a Volcano is Going to Erupt?

As Ask-a-Geologist volunteers, we often get some really interesting questions. At least I call them interesting, anyway, because they open great doors to an interesting geology explanation. One such question follows.

How do you know if a volcano is going to erupt or not? Are volcanoes predictable?  Predictability is an important thing for humankind. If you are being shot at - by an  errant asteroid (like Tunguska, 1908), a hurricane, a tornado, an earthquake, a tsunami, or a volcano - there's some consolation if you can at least predict the event. A warning siren would be nice. This may give you enough time to collect the kids, the dog, the family photo albums, and Aunt Dottie's genealogy list - and beat it out of Dodge.

For the record, we can predict this much:

• Asteroid impact: Something that would destroy a continent, up to 30 years' warning.
• Asteroid impact: a "city buster" 30 to 50 meters in diameter, which could obliterate Washington, DC in seconds: hours of warning - if at all. These are very hard to detect because they are so small. They are so destructive because of the 25,000 km/hour speeds and phenomenal kinetic energy this translates into.
• Hurricane: a week's warning that it is forming, moving towards you... and a day or two warning that you are about to get badly hammered.
• Tornado: hours max, and perhaps as little as 5 minutes'  warning in the Midwest of the United States. And this is with the most advanced Doppler Radar network on the planet.
• Earthquake: no warning - they are still unpredictable. You can at least know if you are in an earthquake hazard zone, and in some parts of California, you can get an actual percentage likelihood that you will get hammered in the next 30 years.
• Tsunami: if you are in Hawai'i and the tsunami is triggered in Chile, then up to 9 hours warning. If you are in Indonesia and the tsunami is triggered by your personal subduction fault, then you get less than 20 minutes warning. Then it becomes: how fast can you run, and how far is it to the nearest high-point? The mayor of Minamisanriku, Japan, had less time than this to get to the communications tower on top of the town hall after the great Tohoku earthquake hit. Wave after wave swept over and gutted the multi-story, steel-framed building, killing all still inside - but he survived with scars on his hands from hanging onto the steel tower. Most people growing up around the Pacific Rim or Indonesian Archipelago are taught this warning from earliest childhood: if the ground shakes, run to high ground as fast as you can. 
• Volcanoes: As much as 6 - 10 months' warning before an eruption - but many of the restive events that trigger warnings end up with a "fizzle" - it goes quiet again. For this reason, the warnings are graded, advanced in stages: Yellow, Orange, Red. If there is going to be a violent eruption, then deformation, gas, and seismic monitoring networks (if installed beforehand) can warn you with "days to weeks", and then as the signals ramp up, with "hours to days" timing, but the magnitude of the eruption is still very difficult to estimate - and with that, its consequences. The destruction of Armero, Colombia, happened about 45 minutes after the first phone-call from up the canyon towards Nevado del Ruiz volcano saying that something "sounding like 10 diesel locomotives" was passing and moving in the direction of Armero. The mayor told people not to worry. The Catholic Bishop told people to go to the cathedral for protection. But NO one can outrun a Lahar. Only the foundations of the cathedral survive. There is an elaborate acoustic flow monitor system down-stream and west of Mount Rainier in Washington State. School children routinely have evacuation drills - they must run a mile to a bridge over a busy highway to get to higher ground. They have just 45 minutes from hearing the first automated siren.

Q:

How do you know if a volcano is extinct or going to erupt again?
Rylee from Mrs. King’s class.


A:
The 169 volcanoes in the United States and its territories are classified by USGS volcanologists as Very High Threat, High Threat, Moderate Threat, Low Threat, and Very Low Threat. There are 18 volcanoes classified as Very High Threat.

These categories were developed after many years of careful mapping, dating, and analysis. They are based on a number of criteria, including the history of the volcano - such as how recently did it erupt? How many times has it erupted in the past 10,000 years? How far out do old eruptive products reach? How many human beings are now exposed to danger in these areas if there is another eruption? Each volcano is sort of like people or bears: they each have their own unique "personalities". Some, like Kilauea volcano in Hawai'i are mostly effusive: they tend to flow lava with little explosive behavior. Others have a long and violent eruptive history, like Yellowstone. 640,000 years ago Yellowstone erupted and laid out a blanket of ash - that ash is over 20 meters (66 feet) thick hear Colorado Springs, CO - over 1,300 kilometers (800 miles) away from the volcano! I have personally pulled a camel's tooth out of the base of that off-white-colored deposit (called the Pearlette Ash Formation) where it had smothered all living things under it.

NO one can outrun a 20-meter-thick blanket of ash that reaches out and covers a continent.

A key point here is that the volcanoes and their eruption products must be age-dated. They must also be carefully mapped to see how far the eruptions reached in the past. It's pretty safe to say that if a volcano hasn't erupted in 10,000 years it's PROBABLY dormant. However, Mount St Helens last erupted in 2004-2006, and before that in 1980-1986. Kilauea volcano in Hawaii has been erupting continuously since 1982, so it's pretty safe to say that these two are DEFINITELY going to erupt again. Those are the two extremes, but a volcano called Four Peaks in Alaska erupted in 2006 after being dormant for many thousands of years... so even apparently dormant volcanoes can surprise us with little or warning - and the warning comes only if they are instrumented. Would YOU spent ~$100,000 to instrument a volcano that last erupted perhaps 10,000 years ago? Something in between Mount St Helens and Four Peaks would be Mt. Edgecumbe near Sitka, Alaska. It hasn't erupted in at least 5,000 years, so it's hard to say if it's extinct or not.

What have we done to protect the American people - and to prevent a volcanic eruption from becoming a volcanic crisis? We have put seismometers and telemetered GPS instruments on almost all of the most dangerous volcanoes. Cleveland volcano in the remote Aleutian Chain (which erupts frequently) is an exception. It has not been instrumented yet because we don't have enough funding to do so - but we watch it daily from satellites. Also, there are no towns nearby, so it was given a relatively low priority. Dangerous volcanoes close to human population centers are all instrumented in some way or another as of this year (2011). This way we will ALMOST always be able to provide some warning, even if only a few days.

When Mount St Helens erupted on October 1, 2004, we had about a week's warning from suddenly increasing micro-earthquake activity. As far as our records show, it was dead silent the day before the first volcano-tectonic earthquakes started (I was standing on the 1980-86 Dome just two months earlier). As a result, the Johnston Ridge Observatory five miles away was evacuated in time and no one was hurt. JRO was named for David Johnston, one of our PhD volcanologists who was killed by the 1980 eruption - when it erupted catastrophically while he was monitoring it. That won't happen again as long as we can keep doing our job protecting the American people.

We are, after all, the United States Geological Survey. Without funding, however, even the most dedicated scientists on the planet are helpless.

~~~~~