The Rumble Underground

A raucous boom. A sci-fi phaser. High-frequency buzzing. These sounds are barely audible in Andrew J. Michael’s office at the Earthquake Science Center in Menlo Park, the seismology center for the United States Geological Survey (USGS) on the West Coast. “Scientists tend to like quiet,” he says, as we hunch over a speaker he’s set up on the floor.

One could be forgiven for expecting that Michael, one of the nation’s leading seismologists, might also prefer quiet. Unassuming with a quick laugh, shirt tucked in, gold wire-rimmed glasses, he has dedicated his life to earthquakes since the day a geology professor visited his junior-high school. His interest led him to MIT, Stanford, two post-doctoral positions, and the USGS, where he’s served as a research geophysicist for the past 28 years.

“I’ll turn it up a bit more,” he says and cranks up the volume. Clicks, knocks, and crackles reverberate off the walls. “I think of the earthquakes being very much like percussion,” Michael says. “They’re not harmonic.”

Describing earthquake sounds is subjective, compromised by experience and the limits of language. That said, they’re ominous. Intuitively, physiologically, we know they embody danger.

Michael points to a seismogram — one of those scraggly EKG charts of the Earth — as he plays the 1992 7.3-magnitude Landers, California, earthquake. When it struck, Landers was the largest earthquake in the contiguous U.S. in 33 years. The recording begins with a rumble so thunderous it overwhelms the other activity captured by the seismogram. “What we’ll hear,” he says, “is that some of these other things — these pulses here in the middle — actually turn out to be small earthquakes.” Each sound reflects the earthquake’s size and depth, the type of fault, the soil conditions — an array of variables.

Michael’s eyes brighten when we listen to the Landers earthquake, this time recorded at Parkfield, a notably active section of the San Andreas Fault. “One thing to listen for that’s interesting,” he says, “that I’ve never really chased down: As it gets out toward the end of the seismogram, you’ll hear a little whistling.” On the third play, I hear the whistle, metallic and hollow, like a subway departing at high speed. It’s a thrilling moment, like learning Morse code or a secret language.

We listen to two recordings of a smaller 1994 earthquake near the Parkfield monitoring station, which captured it, as did another station about 70 miles away, near Hollister. By comparison, the sounds are more attenuated, more individually clear than in the closer recording. With the Hollister recording, Michael turns to his computer and points out a blip of relative calm on the seismogram between the P waves, the first vibrations detected in a seismic event, and the S waves, the second vibrations to arrive. “This is 20 seconds here,” he says. “The farther you get away from an earthquake, the more warning you get, but the less you need it. There’s a sweet spot where you get enough warning to do something.”

There’s just one catch, though. Almost all earthquake sounds are inaudible to humans. On a rare occasion, an earthquake might be big enough and shallow enough that the vibrations will reach the surface at a frequency so high it overlaps with the lower ranges of what we can hear. Most of the time, the sounds we associate with earthquakes are caused by the impact of earthquake activity on man-made materials. “Even in the catalogs of sounds we have,” Michael says, “most of it is not the earth itself — it’s stuff in buildings, around it.”

However, if you have a digital seismogram, which Michael and USGS have in abundance, there’s a fix: Convert the seismogram into an audio file, speed up the file so it runs at a higher frequency, and press play. You’ve got your audible earthquake. This is the purest way to hear an earthquake, Michael says. From a scientific perspective, he emphasizes, these earthquake sounds are considered artificial. Their value is as an educational tool, a way to engage the public about earthquake preparedness and safety. As he says, “Seismologists don’t usually sit around listening to earthquake sounds.”

But sometimes they do. Shortly after the 1992 Landers earthquake struck, USGS Menlo Park held one of its regular open houses. Michael converted Landers seismograms into audio and played them on a boombox. Visitors could adjust the bass and treble to isolate the specific sounds of the earthquake. The sample played over and over, every 30 seconds, annoying some of Michael’s colleagues but captivating everyone else.

A trained trombonist, Michael then took the sounds a step further and composed “Earthquake Quartet No. 1,” which features the Landers and Parkfield earthquakes along with parts for voice, trombone, and cello. The piece is about the “relationship of society to earthquakes.” Since its first performance in 1999, following a lecture Michael gave on “The Music of Earthquakes” at a conference, a rotating crew of noted scientists has performed the quartet, piquing the interest of experimental musicians worldwide with its haunting discordant tones.

It’s probably a good thing we can’t hear the earth shifting under our feet. Our nervous systems would be in constant fight-or-flight overload. But listening to earthquake sounds in a controlled and edified manner — as music? Pure puckish thrill.