Listening for a Tornado’s Infrasound Roar May Provide Better Warning Systems

In May of 2013, an EF5 tornado—the most powerful class—devastated the city of Moore, Oklahoma and the surrounding area, killing 24 people and wounding more than 200. The tornado leveled entire blocks of houses, destroyed schools and medical buildings, and tossed cars around, wreaking havoc on the city.

The weekend before the storm, engineering professor Brian Elbing and his wife had flown to Oklahoma and purchased a house, in preparation for his new job at Oklahoma State University (OSU). “This certainly gave us a pause on our decision,” says Elbing.

Five years later, Elbing is still living in tornado alley and, he says, “the longer we live here the more comfortable we are.” That’s not to say he’s forgotten about the very real dangers tornados pose. In fact, Elbing now applies his professional expertise in fluids and acoustics to study them, analyzing the sound produced by storms in hopes of better understanding tornado formation.

As illustrated by the traffic jams that lined the streets of Moore when the tornado hit, current warning systems don’t provide people enough time to get themselves and their loved ones to safety. Advances in this area have progressed slowly and incrementally, especially in the last 15 years. Elbing thinks it’s time for something new.

At the 175th Meeting of the Acoustical Society of America (ASA) earlier this month, Elbing reported the results of new research into the sound produced by tornados. Performed in collaboration with OSU graduate student Christopher Petrin and the University of Nebraska’s Matthew Van Den Broeke, this research could bring us closer to more reliable warning systems.

In the last 20 years, scientists have demonstrated that many geophysical events produce infrasound, sound below the threshold of human hearing (20 Hz). In fact, there is evidence that tornado-producing storms emit infrasound signals more than an hour before a tornado actually forms. The hope is that these signals contain enough information to trigger reliable tornado warnings on their own, or to indicate that meteorologists should send drones to an area to investigate a possible funnel cloud forming.

While the aim is straightforward, the process is not. One reason is that if you could hear in the infrasound range, you’d probably wear earplugs. “One of the biggest advantages of infrasound is that it dissipates slowly in the atmosphere, allowing it to be observed well beyond line-of-sight,” says Elbing. “However, this is also the primary disadvantage for studying infrasound. While it is great we can hear our source of interest from so far away, we hear everything else too,” he says.
All kinds of things produce infrasound—natural sources like whales, alligators, earthquakes, and waterfalls, as well as human-created sources like explosions, sonic booms, diesel engines, and wind turbines.

Infrasound research took off in the 1970s as a way to monitor nuclear blasts, but nuclear blasts emit infrasound below 0.5Hz. Sounds in range of 0.5 Hz to 20 Hz –like those produced by tornados—were largely ignored until recently, so research is still in the early stages. Before scientists can unravel the information contained in the infrasound signals of tornado-producing storms, they need to find out whether they can even untangle the sounds of tornados from the infrasound background noise.

To this end, OSU members of CLOUD-MAP, a multi-university collaboration that aims to improve weather monitoring and forecasting, installed a system of three microphones on campus in January 2017. Designed for long-term infrasound monitoring, the system is arranged so that the microphones form the points of triangle with sides about 200 feet long. This pattern allows researchers to extract information about the location where a signal originates.

Four months in, the microphones experienced their first tornado. A storm produced a small EF0 tornado about 12 miles from the monitoring station on May 11, 2017. As presented as ASA, Elbing and his colleagues analyzed the infrasound recorded before, during, and after the storm, in order to see what they could learn about the tornado signal. By comparing the infrasound recordings to one another and to theoretical estimates for the noise produced by the tornado, the team determined that the tornado did, in fact, produce infrasound that was captured by the microphones.

Measuring a clear signal is valuable progress, but it’s not the end goal. “We know that tornadoes emit infrasound, that it can start as much as hours beforehand, that infrasound carries information about its source, and that infrasound can be detected from hundreds of miles away,” says Elbing. “What we don’t know is how to interpret these infrasound signals.”

That’s what makes this research so important. Currently, there are just a few clear measurements of infrasound produced by tornados, and we don’t even know yet how that sound is produced. Knowing the source of the sound is key to connecting infrasound signals to the properties of tornados. That’s where Elbing’s interest lies, in developing models of sound production mechanisms and testing them against observations. But this requires more observations.

Living in tornado alley has its downsides, but there is a silver lining: New data is never far away.

—Kendra Redmond