Seismic Shift

If a nuclear sub blows up in the ocean and no one hears it, did it really explode? The Russians won't say, but a seismologist in Arizona knows the answer

By Dana Mackenzie, Dan Winters|Saturday, September 01, 2001

The world's most advanced seismometer, a Streckiesen STS1, uses electronic feedback to measure vibrations of as small as a few micrometers and as brief as a millihertz. A bell jar isolates the mechanism from atmospheric pressure changes.
It's a small world, people say, but they have no idea how small until they've been in Terry Wallace's office. On a recent morning, the prickly pear is budding in Tucson, the ocotillo is in full bloom, and Wallace is eavesdropping on a seismic disturbance 8,000 miles away. His computers at the University of Arizona simultaneously monitor 500 seismic stations around the world over the Internet, and they can link to hundreds more if there is a special reason to— as there is this morning.

A report has appeared in The Washington Times that the Chinese may test a small nuclear bomb at Lop Nur, in western China. So Wallace has connected the terminal in his office to an array in Kyrgyzstan, a former Soviet republic that borders China, just in case something happens.

Around 10:30 in the morning, something does. A hefty tremor sets the lines on Wallace's screen wiggling.

"We've got an event here," Wallace says, a trace of incredulity in his voice. Within minutes, he has a map up on his computer screen showing a bright red dot on the Afghanistan-Pakistan border. "The event we just saw occurred down here," he says. "It's about magnitude 4.5. It's definitely an earthquake." Not an explosion, not an international incident, but nevertheless a sizable jolt, one that plenty of people in eastern Afghanistan noticed.

And one person in Arizona.

For Wallace it's a routine event, soon forgotten. As this country's leading seismic detective, he is less interested in earthquakes than in man-made jolts: airplane crashes, industrial explosions, terrorist bombs. Most of the time they register fairly low on the Richter scale, and companies or governments often don't want anyone to know about them. But chances are that Terry Wallace will find out anyway. On a planet as small and as wired as ours, it's hard to keep a secret from a good seismologist.

This August marked the first anniversary of Wallace's most important case, a disaster that made the front pages of newspapers around the world. For Wallace it began with a quiet early-morning look at the previous day's seismograms. Seismic events not associated with earthquakes turn up every other day or so. Their sources range from meteorite impacts to sonic booms from the space shuttle, but it's difficult to sift them out from Earth's natural creaks and groans— about 200 events of magnitude 4 and 2,000 events of magnitude 3 per week. To narrow the field, Wallace usually starts by looking at known hot spots— say, within 100 miles of a known nuclear test site.

On Saturday, August 12 of last year, a tremor from the Barents Sea, 90 miles north of the Arctic Circle, reached seismic arrays in Norway, Finland, and Russia. Wallace, when he first saw the record of the tremor on Monday morning, ignored it: The Barents Sea was too calm, politically, to merit his attention. But 5,000 miles away, seismologist Frode Ringdal took another look. According to the morning news, the Kursk, a powerful Russian nuclear submarine, had been crippled by an accident during a military exercise. Now it was stranded on the floor of the Barents Sea, beneath 350 feet of icy water, with 118 sailors aboard. Ringdal, who is scientific director of the NORSAR Data Center in Kjelling, Norway, began looking immediately for telltale seismic signals. Right away, he spotted an unusual burst of waves from the Barents Sea, he says. Still, the timing of the Barents Sea tremors didn't match the reports, which said that the accident had taken place on Sunday rather than Saturday.

On October 31, 1989, seismic events of three kinds were recorded in Tucson, Arizona. Top: An ammonium-nitrate-based explosion, with the strength of 50,000 pounds of TNT, in a copper mine south of Tucson. Middle: A nuclear explosion, equivalent to 20 to 100 kilotons of TNT, at the Hornito test site in Nevada. Bottom: An earthquake in southern California measuring 4.4 on the Richter scale.
In Russia, the sailors' families wanted to know whether their sons and husbands were alive. In neighboring countries such as Norway, people wanted to know whether the sub's two nuclear reactors had exploded or released radioactive material. But Russian officials dithered and stonewalled. President Vladimir Putin, vacationing on the Black Sea at the time, apparently decided the incident wasn't worth interrupting his break. Meanwhile, officials floated a variety of scenarios. They claimed that contact had been established with the crew, which had 72 hours of oxygen left. They speculated that the Kursk had run into an old World War II mine, run aground, or collided with an unidentified foreign submarine.

The foreign-sub theory conveniently pinned the blame on old cold-war enemies, and Russian officials stuck to that story for weeks. As late as October 29, Admiral Vladimir Kuroyedov, the Russian navy's commander in chief, said he was "80 percent certain" the accident had been caused by a collision, and he promised to name the "killer" by November 5. That day came and went without an announcement, but on November 8, Deputy Prime Minister Ilya Klebanov reiterated that there was "serious visual evidence" of a collision, including a large dent and long grooves in the side of the sunken sub.

Between times, before President Putin even returned from his vacation, Wallace had set up a Web site devoted to seismic analysis of the Kursk accident. Hundreds of Russians, desperate for information, linked to it. And within a week of the accident, Wallace had figured out what had really happened.

When Wallace first heard about the accident on August 14, he called up seismographic data from NORSAR and stations in Russia and Finland. Seismic activity is rare in the Barents Sea, and no other tremors had been recorded that weekend. So the Saturday blast— which Ringdal first estimated at magnitude 3.5 and Wallace later revised up to 4.2— seemed as conspicuous as a car alarm. The signals had arrived at the various seismic stations at different times because each was a different distance from the blast, so Wallace was able to triangulate the source. It was exactly where the Kursk was reported disabled.More secrets soon emerged. As Wallace synchronized and combined seismic waves from five stations— a process called beam forming— random sounds canceled one another out, and authentic seismic signals reinforced one another. Little by little a precursor event stood out from the background. It occurred 135 seconds before the main catastrophe and was 1/250 as large. But the two events made nearly identical seismic waves.

When something explodes underwater, it creates a bubble of hot gases. As the bubble rises to the surface, it becomes larger and smaller several times per second. That creates a pattern of sound waves known as a bubble pulse. From years of experience in offshore-oil prospecting and nuclear-bomb testing, seismologists have become quite skilled at detecting bubble pulses.

The signal seismometers picked up at 11:30 that Saturday morning was a perfect example of a bubble pulse— like a thunderous chord played by a giant orchestra. The lowest note in the chord, at 1.5 cycles per second, was seven octaves below middle C— far below the threshold of human hearing. But the seismometers heard it loud and clear. Nothing else in the world— not a collision, not an earthquake, not even a meteorite from outer space— could have produced that signal. The Kursk had gone down in an explosion.

Eventually, Wallace combined seismic evidence with press reports to sketch out the Kursk's final moments. Together with Keith Koper of the University of Arizona and Stephen Taylor and Hans Hartse of Los Alamos National Laboratory, Wallace published an analysis this year in Eos, the journal of the American Geophysical Union.

The explosion of the Kursk sent seismic ripples across the Barents Sea and into Scandinavia, where more than 20 seismic stations picked them up.
Graphic by Matt Zang
Based on the Data of Keith D. Koper and Terry C. Wallace/University of Arizona, and Stephen R. Taylor and Hans E. Hartse/Los Alamos National Laboratory
Seismograms from the ARCES station in northern Norway show the ground shaking in three directions. From top: up and down, north-south, east-west. Measurements are in centimeters per second.
On the morning of August 12, the Kursk had been cruising near the surface. Shortly before the accident, some sources say, the submarine sent a coded radio message to its command ship, Peter the Great, apparently requesting permission to fire a torpedo or a cruise missile. The reply came back: dobro, or "good." That was the last message from the sub. A few minutes later, either the torpedo or its volatile liquid propellant exploded with the strength of 150 kilograms of TNT, producing the smaller of two seismic pulses.

The next two minutes must have been chaos aboard the Kursk. The radio and sonar center, located in the third of nine compartments, seems to have been destroyed immediately, because no distress signal was received. Fire raged through the front sections of the sub. The hull had been breached, and water poured in. With its power crippled, the submarine nosed gently downward. Then things got worse.

When the sub hit bottom, at a depth of between 280 and 350 feet, several torpedoes or their propellant exploded at once with the force of three to seven tons of TNT. Wallace assumed at first that the entire crew must have perished instantly: "If you were 300 feet away, the pressure was large enough to blow your skin off," he says. But when Russian divers finally boarded the sub in late October, they found that some sailors in the ninth compartment had survived the blast, presumably because this compartment had been sealed off before the larger explosion. One sailor, Lieutenant Captain Dmitri Kolesnikov, had scrawled a message in the dark indicating that 23 men were still alive four hours after the accident. Early Russian reports stated that crewmen were alive and in contact with the surface as late as August 14, but Wallace disagrees. "We know for a fact from the notes that these other guys died long before the 14th," he says. "They would have died within a day from asphyxiation." The original report, which claimed that the crew had 72 hours of oxygen, assumed that the sub had sunk intact. The sealed-off section was contaminated by fumes from the fire, and sailors there would have run out of breathable air in a short time.

Some questions remain, such as what type of torpedo caused the disaster. Western military experts believed it might have been a new experimental weapon called the Squall. More recent reports suggest it was an old, outmoded torpedo and that the crew may have even known it was defective. But few people dispute Wallace's explanation of the accident.

"I completely agree with this," says Viktor Baranets, a former army colonel who now writes for the Moscow newspaper Komsomolskaya Pravda. "I have worked directly with the navy, met with the commission, with the commander of the naval fleet, with scientists, and a long list of others. They have studied the question thoroughly and from all sides. One must consider the cause to be an internal explosion." Baranets doesn't expect the commission will issue a final report until after the Kursk is raised. The recovery, which was scheduled to take place this summer, has been delayed because the government lacks the funding. In the meantime, Baranets says, "time is working against the truth."

Since last September, the Russian navy has regularly set off depth charges around the sunken sub to discourage inquisitive visitors. Ironically, the explosions have helped Wallace pinpoint the location and size of the Kursk's explosions. Just as CAT scanners produce three-dimensional images of the body by irradiating it from different directions, seismic arrays, measuring signals coming in from various directions, can reconstruct a three-dimensional view of an area's geology. For example, seismic waves travel more rapidly through granite than through sandstone and arrive at the array sooner. By compensating for this effect, seismologists can precisely estimate the distance to the source of the signals. "On some days, we can locate [the depth charges] in a straight line," Wallace says, "which suggests they were set by a boat going along a grid pattern. We can tell you the speed of the boat. That's what forensic seismology is all about."

Terry Wallace and a classic Benioff seismometer, now retired. In the background, portable seismometers, stacked on cabinets filled with historical seismograms, await field duty.
Wallace hardly looks the part of a detective, with his open face, sandy hair, and rimless glasses. But his interest in both geology and explosions comes from a long family tradition: His great-grandfather was a prospector, his father a chemist who moved to Los Alamos to work on nuclear weapons.By the time Wallace was in tenth grade, he knew he was going to be a seismologist— if he didn't become a mineralogist first. Moving to Tucson 18 years ago allowed him to satisfy both interests. Tucson hosts the world's largest annual show of gems and minerals, where Wallace indulges his passion for silver-bearing minerals.

The same passion drives Wallace's seismic investigations. Every night, his computer search engine scours 22 news sites on the Web for six telltale words: explosion, nuclear, seismic, meteorite, landslide, and bolide (a fireball). Every morning, he tries to match news stories with seismic reports. Sometimes patterns emerge. "The last two years have been tough in China," he says matter-of-factly. "There have been fireworks factories blowing up." These have been a major embarrassment to the Chinese government, which has been less than completely forthcoming with details. The last explosion, in March, killed 42 children and teachers in a school in Jiangxi Province. Wallace was tipped off to it by the BBC's Web site. Only a week after the Kursk explosion, Wallace investigated a deadly natural-gas explosion in New Mexico [see "Great Ball of Fire," left], and most recently he helped locate where two American F-15C fighters crashed in Scotland on March 26. (The Air Force didn't ask for his help, Wallace says: "No one would have thought of asking a seismologist.")

"It's public-interest seismology," says Anne Paquette, an undergraduate student who helped gather the data on the depth charges around the Kursk. But like all good science, it doesn't take sides. A few years ago, Wallace and other seismologists embarrassed the U.S. intelligence community, which had accused the Russians of violating the Threshold Test Ban Treaty by setting off a nuclear explosion that was larger than allowed. Wallace's analysis showed that the suspect signal was really an earthquake 50 miles offshore from the Russian test site. Not long thereafter, Wallace published an article showing that an explosion the Indian government claimed to have set off with a thermonuclear device was, in fact, only a conventional nuclear explosion. "I got literally hundreds of letters from Indians," he says. "I was this imperialistic American who was questioning their ability to build nuclear weapons."

"Terry sticks his neck out a little bit," says George Zandt, a colleague at the University of Arizona. "He'll put out a possible solution prior to when a government agency would like. There's a running friction between him and government agencies who would rather not have university seismologists fiddling in their field." Wallace does this partly to prove a point: that governments cannot and should not keep information about man-made seismic events a secret. And he does it partly because, as he says, "I am a scientist— we just want to know what happened."

In the meantime, the Internet grows, and an increasing number of seismic stations use it to share their results openly. Before long, it might not take so much skill to interpret the readouts. "I see the day 10 years from now," Wallace says, "when any news organization will have the same seismic-signal software as I do now, the same virtual seismic network, and anybody in the world will know immediately when there's an earthquake." Or anything else that shakes the Earth.

The Nuclear Watch

It might seem odd that there are so many seismic stations in the remotest areas of Scandinavia, but there is an explanation: The island of Novaya Zemlya, where the Russians test their nuclear weapons, is just across the Barents Sea from the region.

Many of these cold-war-era stations are now part of the International Monitoring System, a global network that was set up to monitor nuclear blasts for the Comprehensive Nuclear Test Ban Treaty. The treaty, which bans nuclear explosions of all kinds, lives in a complicated diplomatic and legal limbo. Signed by 161 nations and ratified by 76, it will have no legal force until the United States and 12 other "Annex II" nations ratify it. And the United States won't do so anytime soon: The treaty suffered a stinging defeat in the U.S. Senate two years ago.

Still, the global network envisioned by the treaty is in place, financed by the United States and other nations. An international data center is up and running in Vienna, and information from seismometers worldwide is shared openly. Even five years ago, Terry Wallace's rapid monitoring of seismic disturbances worldwide would have been inconceivable.

Opponents of the treaty, such as Arizona senator John Kyl, argue that an absolute ban on nuclear testing cannot be enforced and that it would be too easy for cheaters to escape detection. To Wallace, the real issue is not whether small, individual blasts can be detected but whether a country could mask a series of 1-kiloton blasts of the kind needed to develop a new nuclear weapon. He regards the Kursk submarine case as proof of the system's efficacy. Although it was much smaller than a typical nuclear test, the Kursk explosion had about the same seismic magnitude as a 1-kiloton nuclear blast on land because it happened underwater. Water efficiently transmits seismic energy.

"Whoever wants to evade detection will have to avoid a hodgepodge of stations all over the world," Wallace says. "It's like a global neighborhood watch."
— D.M.

Great Ball of Fire

On August 19 last summer, two families camped at a popular fishing spot on the Pecos River, near the New Mexico-Texas border. At 5:26 a.m., a 30-inch-wide underground gas line— one of eight that supply the West Coast— exploded nearby, gouging out a crater 86 feet long, 46 feet wide, and 20 feet deep. A huge fireball, visible from the city of Carlsbad 20 miles away, engulfed the campsite. The temperatures were so high that rescue personnel couldn't reach the local shutoff valve, and the fire burned for almost an hour before the El Paso National Gas Company turned off the gas flow farther upstream. All 12 people at the campsite died.

Afterward, two nearby seismic arrays not only pinpointed the exact time of the rupture, but they also showed that there had been three explosions, not one. The first and smallest explosion blew out the crater, but there were no flames. "It must have been a terrifying thing," says Rick Aster, a seismologist at the New Mexico Institute of Mining and Technology who set up one of the arrays. "These people had 24 seconds to run for their lives." Then something ignited the gushing gas— perhaps a spark from a campfire, or the ignition spark from the truck that two campers used to flee. The third explosion (the tallest spike on the seismogram above) occurred 19 seconds later and may have been due to a second rupture in a nearby section of the pipeline. For the next 58 minutes, the seismometer recorded an eerie, earthshaking sound: the uninterrupted roar of one of California's eight natural-gas supply lines going up in flames.

The seismograms may yet be introduced as evidence in a court case. "My impression is that they are still trying to negotiate," Aster says. "They know they have a potentially very damaging lawsuit on their hands." One of the first questions might be why they took so long to turn off the gas. — D.M.
— D.M.

The U.S Department of Energy Web site for its Nuclear Explosion Monitoring Research & Engineering Program provides technical information about monitoring technology:

For information from the Lawrence Livermore National Laboratory on how forensic seismology would support the Comprehensive Test Ban Treaty, see

A detailed time line of events relating to the sinking of the Kursk is available at

Comment on this article