A tornado decimated parts of Tuscaloosa, Ala., in April 2011. (Dusty Compton/AP)

Technology developed to detect nuclear weapons detonations from afar might help meteorologists track tornadoes.

A team of scientists are hoping to enhance tornado detection with some help from an unlikely source — specialized microphones developed to assist governments in spotting illegal nuclear weapons tests.

Tornadoes are one of the most challenging extreme-weather events for forecasters to predict and warn against. Human eyeballs remain one of our primary tools for confirming these violent vortices, often too small to leave a clear, unequivocal signature on radar.

But Roger Waxler of the University of Mississippi’s National Center for Physical Acoustics believes tornadoes may be more trackable from a distance by listening to them.

Waxler’s team, which previously developed a state-of-the-art microphone for picking up infrasound, or sound below the range of human hearing, presented evidence at the American Meteorological Society meeting in Boston last month that tornadoes emit infrasonic signals that can be heard more than 50 miles away.

More work is needed to determine whether tornadoes have a distinctive sonic signature. If that turns out to be the case, it’s possible infrasound microphones could one day augment current tornado warning systems, which are mainly based on Doppler radar and storm spotters.

“If this idea does work, it could be useful for forecasters to be situationally aware that a tornado has formed,” said Brice Coffer, an atmospheric scientist at North Carolina State University. Coffer isn’t involved in Waxler’s work.

To determine whether a storm is likely to spawn a tornado, meteorologists look for several key ingredients, including high vertical wind shear in the storm’s lower levels and atmospheric instability, which promotes the development of strong updrafts. They also track the storm via radar, keeping an eye out for telltale signs such as a curling of the precipitation field — a hallmark of tornado-producing supercell thunderstorms — and rotation inside the storm.

Doppler radar products can also detect the debris lofted by a tornado touchdown, offering radar confirmation of a twister.

If the right ingredients are present, forecasters will issue a tornado warning. But because tornadoes are meteorologically minuscule, actual detection of one very often depends on witness accounts. This is especially problematic in the southern United States, where a combination of topography and trees often makes it hard to see tornadoes coming — part of the reason the region has some of the highest tornado fatality rates in the country.

Waxler believes infrasound could act as another pair of eyes (or ears), beefing up real-time tornado detection and giving communities precious extra minutes to prepare.

“We’re hoping we could save lives,” he said.

Waxler’s group started developing infrasound microphones in the early 2000s to help the U.S. government enforce the Comprehensive Nuclear Test Ban Treaty, which relies on a global network of seismic stations, microphones and radionuclide detectors to spot illegal nuclear tests. By 2010, they were field-testing the sensors at a large scale.

Concurrently, the group received funding from the National Oceanic and Atmospheric Administration to study infrasound monitoring of violent weather. Working with Hyperion Technology, the group deployed infrasound arrays around Oklahoma to listen for tornadoes. The initial results were promising.

During the past few years, with additional funding in support of NOAA’s VORTEX Southeast campaign, Waxler’s group has continued this research, deploying a network of infrasound microphone arrays across northern Alabama. On March 19, 2018, a storm system passed over the region, spawning at least eight tornadoes that the team’s arrays picked up.

These tornadoes, which were the focus of Waxler’s recent AMS presentation, add to a growing pile of recent evidence that tornadoes produce infrasound signals in the 1-10 Hertz range. According to Waxler, the team’s recent work also shows that signals from tornadoes can be separated from certain types of background noise.


Doppler radar reveals the powerful rotation associated with a Dallas tornado in 2019. (GR2 Analyst/Matthew Cappucci)

“The meteorology community is concerned we’re picking up sounds that have nothing to do with tornadoes,” Waxler said. “So far, we’re showing that’s not true. We can cleanly differentiate thunder [and] stationary signals, probably factories.”

While the results are intriguing, there’s still a lot of work to be done.

According to Paul Markowski, a meteorology professor at Penn State who’s collaborating with Waxler’s group, there appear to be multiple sources of infrasound coming from inside storms. To be useful as a real-time tornado tracking tool, tornadic infrasound needs to be distinguishable from all the rest.

“[I]n scientific experiments we ordinarily have a control case to compare against,” Markowski wrote in an email. “I haven’t seen what infrasound looks like from these same sensors near storms where we unequivocally know there are no tornadoes.”

Matthew Parker, an atmospheric scientist at North Carolina State who isn’t involved in Waxler’s work, agreed the possibility that non-tornadic storms produce the same signals needs to be much more rigorously ruled out, as well as the possibility that not all tornadoes produce infrasound.

“It seems that some tornadoes do emit infrasound, but that non-tornadic storms also emit infrasound,” he wrote in an email. “It is incumbent on the community to understand what are the actual sources of infrasound in storms, and whether there are signals that are truly distinctive to tornadoes.”

Waxler is hopeful the open questions about tornadic infrasound can be overcome with additional acoustic monitoring of a variety of storms. The team has NOAA funding to collect more field data this year, and it’s expanding its monitoring network to cover parts of Mississippi and Louisiana, as well.

Another question is why tornadoes produce infrasound at all — is the signal coming from vibrations within the vortex, the release of heat as water vapor condenses, or something else? Nobody’s sure, but since 2017, physicist Brian Elbing of Oklahoma State University has been recording the atmosphere around campus with three infrasonic microphones in an attempt to find answers. His team is building a second array that it is hoping to deploy in a new location by the spring.

If the mechanism responsible for producing these signals can be identified, scientists might be able to use infrasound to infer other properties of a tornado, he said.

“The key is answering these fundamental questions,” Elbing said. “With more mics out recording, we’ll learn more.”