Lightning flashes through a cloud of ash pouring out of a remote Alaska volcano. A radio signal from the electrical discharge travels thousands of miles in milliseconds and gets picked up by dozens of antennas. A computer sends an alert to volcano researchers.
The volcanologists warn pilots, and a plane traveling from Anchorage to Asia -- with engines that will be choked out by the ash -- steers clear of the dangerous plume. It's a disaster averted.
At least, that's the theory, said Stephen McNutt, a research professor with the Alaska Volcano Observatory and the University of Alaska Fairbanks.
McNutt says lightning detection might be the best way to confirm an explosive eruption at distant locations that lack better instruments. That would include Cleveland Volcano on an Aleutian island about 940 miles southwest of Anchorage and close enough to international air shipping routes to make pilots and volcanologists nervous.
During its most recent unrest, lava has slowly oozed inside Cleveland's crater since July, suggesting a possible explosive eruption to come, according to the Alaska Volcano Observatory. A growing lava dome reached the crater's rim this month. So far, it has apparently not toppled over the edge, increasing the possibility of an explosion, the observatory said.
Unlike Redoubt Volcano, which is much closer to population centers and most recently erupted in 2009 after months of rumbling, Cleveland has no instruments like seismometers and microphones, used to sense magma moving within the volcano.
The closest instrument to Cleveland, a lone seismometer, is in the village of Nikolski about 45 miles away. A larger network of instruments is about 80 mile away at Okmok Volcano, McNutt said.
"Occasionally, in the past, some of the stronger eruptions from Cleveland, the signals have been strong enough to record on these other instruments. But it hasn't been very consistent," McNutt said.
Scientists don't know if Cleveland's dome growth continues or has merely slowed in the past week. But without seismometers on the volcano to feel its inner rumblings, the observatory said it might take hours for satellites to detect ash from an eruption. So for the last week or two, the volcanologists have been using a global network of lightning sensors operated by the University of Washington to keep tabs on Cleveland, McNutt said.
"Cleveland's been on the list of volcanoes to monitor for 10 years or more, but the finances have simply not allowed it," McNutt said. "We had a series of discussions about what could we do that we're not doing, and lightning was one of the few that we could do right now without a huge effort."
Enter the World Wide Lightning Location Network.
The network is composed of more than 50 sensors around the world that can pick up signals from most large lightning strikes, said the network's director, Robert Holzworth. More recently, the network has been used to specifically monitor near volcanoes, also referred to as caldera, Holzworth said.
"Every minute, we look for lightning over 1,562 caldera around the world," Holzworth said.
The network currently monitors 262 volcanoes circling the Pacific Ocean, including Cleveland, for the Alaska Volcano Observatory, Holzworth said. Scientists like McNutt are still studying the effectiveness of using the system to detect eruptions, as well as the causes of volcanic lightning, Holzworth said.
"It doesn't seem like it always happens, but it happens a fairly high percentage of the time, especially as eruptions get larger," McNutt said.
Right when the eruption occurs, rock inside the volcano breaks apart, McNutt said. That fracturing causes an electrical charge to build up on ash particles as the ash shoots up through the volcano, he said. The positive and negative charges get separated, and water vapor turns to ice high in the column, McNutt said. Charge separation, aided by the ice, causes a huge electrical discharge: lightning, he said.
"It's effectively like a dirty thunderstorm," McNutt said.
When lightning strikes, out goes the radio signal, and the network's stations around the globe pick it up as far as 12,000 kilometers away, Holzworth said.
If the signal is strong enough, the network "grabs" a little bit of data -- 1.3 milliseconds worth, to be exact, Holzworth said -- and sends it to a lab in Washington. A computer there uses multiple data points from different antennae to triangulate the location of the lightning, he said.
Then, in a separate calculation, another computer program looks at the lightning's location in relation to a volcano, Holzworth said. The computer counts the number of lightning flashes within two perimeters to the volcano: one, a 20-kilometer circle, the other a 100-kilometer circle. The outer ring is a "veto" circle, Holzworth said.
"If there's a lot of lightning around the volcano, out at large distances, then we'll say, 'OK, looks like natural cloud lightning that's causing the problem,' " Holzworth said.
But if the lightning is detected inside the smaller perimeter, closer to the volcano, and not in the larger perimeter farther away, the system triggers an alert, he said.
"It basically says, 'Hey, go take a look at that volcano because we're getting a lot of lightning right over the summit, which may be from an eruption," Holzworth said.
The alerts come in the form of emails and text messages, said Holzworth, who gets about 10 such texts a day, he said. The alerts have to be taken as just another warning, another tool to help keep skies safe, Holzworth said.
"It's not an idiot sensor. It doesn't just say, 'Yep, there's an eruption going on or not.' You have to have a little bit of capability to look at the clouds or pull up a satellite image so you know what you're looking at," Holzworth said.
And McNutt said seismometers and specialized microphones are much better ways to predict ahead of time if a volcano might erupt. An ash plume can take up to 10 minutes to get to the altitude where it can produce lightning, McNutt said. If it's the only warning of an eruption, that may be too late, he said.
"By the time that's occurred, the ash is already up and in airways where the airplanes are," McNutt said. "Airplanes fly about five miles per minute. In that 10 minutes, an airplane can go from safety to difficulty. So it's not trivial."
While the technique is still being tested, lightning detection appears to be better than no warning at all, McNutt said.
One thing the volcanologist knows for sure is that it never gets old looking at pictures of lightning shooting through an ash plume, especially when the photos show up on the national news, as a picture in May from a volcano in South America, he said.
"They showed volcanic lightning on TV, and I was like, 'Yes! We have arrived.' "
Reach Casey Grove at casey.grove@ adn.com or 257-4589.