Listening for turbulence
By Keith Button|April 2023
In recent months, airline passengers have been injured by encounters with clear air turbulence that, at the moment, crews have no way to detect. A solution could be at hand through a collaboration between NASA and the National Research Council Canada. Keith Button tells the story.
The reference to the death of a passenger on a business jet in March has been updated to reflect the U.S. National Transportation Safety Board’s preliminary report.
Hawaiian Airlines flight 35 was flying through the bright morning sky above a cloud layer about 40 minutes from Honolulu when something unexpected happened. “A cloud shot up vertically (like a smoke plume) in front of the airplane in a matter of seconds, and there was not enough time to deviate,” according to the preliminary report about the December incident. The captain alerted the lead flight attendant about possible turbulence, but it was too late: Almost immediately, the plane was buffeted severely. One passenger was flung in the air with enough force to dent the ceiling. Overall, 25 people were injured, including six with serious injuries, according to the preliminary report from the U.S. National Transportation Safety Board.
The Hawaiian flight was one of a pair of such incidents. In March, seven aboard a Lufthansa jet over Tennessee were injured badly enough to be hospitalized. A third incident involving a business jet was initially blamed on severe turbulence, but NTSB later attributed it to a control malfunction.
Why are such accidents all too common? A plane’s onboard weather radar can easily spot possible turbulence, provided the air ahead contains water droplets, ice crystals or dust. Sometimes, though, turbulent air lacks such markers. In the Hawaiian Airlines flight, the onboard weather radar showed no returns, according to NTSB. Reports from aircraft ahead are helpful, but not foolproof. Someone has to be first to encounter turbulence, and in the Hawaiian case, there was no such warning.
Now, however, researchers are making progress on a possible solution. Clear air turbulence occurs most often when a cooler jet stream flows on top of a layer of warmer air and the rising warm air collides with the sinking cool air, producing what Canadian research pilot Anthony Brown describes as a horizontal tornado called a vortex tube that sends inaudible low-frequency sound waves racing through the atmosphere. If aircraft could be equipped to hear these infrasonic waves, the result could be an early warning system in the cockpit, perhaps supplemented by ground microphones and detections from other airliners.
Researchers at NASA and their collaborators at the National Research Council Canada, including Brown of the council’s Aerospace Research Centre, are in the midst of flight testing a microphone that they believe could do the job. One challenge will be finding clear air turbulence to see how the microphone does. If things go as planned, someday an array of these silver, cylinder-shaped mics could be affixed in the wingtips and fuselage of an airliner to listen for turbulence.
“We should be able to pick up these clear air turbulence signals from a certain distance ahead of time so that pilots have enough warning,” says Qamar Shams, the principal investigator for the infrasonic research at NASA’s Langley Research Center in Virginia.
After this article was reported and written, NASA informed us that Qamar Shams died on March 18 of a sudden cardiac arrest. We have elected to keep the article in its original form to retain the spirit of his interactions with Keith Button and the editing staff.
Additionally, airlines could also chart efficient routes around clear air turbulence, rather than giving a wide berth based on reports from pilots as is done today. Then there is the matter of wake turbulence caused by the trailing vortices from the wingtips of the aircraft ahead. Like clear air turbulence, trailing vortices are invisible to radar in the absence of precipitation or dust. Because they can’t be seen but can persist for 50 kilometers, air traffic controllers require planes to maintain long gaps as a precaution.
If a solution is indeed at hand, it’s one that’s required decades of brainstorming, creative research and persistence through personal loss.
In 2001, Shams and Allan Zuckerwar, his research partner at NASA Langley, were casting about for topics to study. They were intrigued by an article in the March 2000 issue of Physics Today, “Atmospheric Infrasound,” co-authored by NOAA research scientist Alfred Bedard, that described potential applications for infrasound detection, including listening for turbulence with microphones. Shams, an electrical engineer who emigrated to the U.S. from Pakistan in the early 1990s, and Zuckerwar decided to begin early research, thinking they might win funding for a project to demonstrate detection of clear air turbulence and trailing vortices.
As Shams recalls it, in 2003 or 2004 he and Zuckerwar received the go-ahead from NASA, with Zuckerwar initially in the principal investigator role.
They began a set of experiments with off-the-shelf scientific microphones. Locating clear air turbulence would require determining the direction the sound came from, so in a ground test they employed the microphones in triangulation: Three were installed on the ground as the points of the triangle, 30 meters from each other. The direction of the incoming sound was calculated by the milliseconds difference in the time that each microphone detected the sound.
They also bolted a fourth mic atop the giant steel gantry frame at Langley’s Landing and Impact Research Facility where Neil Armstrong practiced the moon landing. “That idea didn’t work,” Shams says, because the microphone picked up too much noise from the gantry vibrating and swaying in the wind.
During two years of continuous testing, the results were encouraging. Whenever a pilot in the area reported an encounter with clear air turbulence to air traffic controllers, the mics detected the same turbulence, out to a radius of about 560 kilometers.
Still, Shams and Zuckerwar concluded they needed even better microphones. Although the commercial mics could pick up infrasonic frequencies, they were meant for detecting audible rather than infrasonic frequencies. So, they built their own prototype mics with larger diameters and larger chambers, and eliminated the air vents in the off-the-shelf design. With these modifications, they discovered they could detect infrasonic sound waves with much better sensitivity. They hired a contractor, PCB Piezotronics in Depew, New York, to manufacture the microphones.
These new versions proved to be nearly 10 times more sensitive than the off-the-shelf mics. From Langley, they detected a rocket launch in Florida 1,050 kilometers away and variations in nighttime traffic through the Lincoln and Holland tunnels of New York City 460 kilometers away.
Armed with the highly sensitive mics, they set out to build a library of sound signatures, meaning the loudness and frequency attributes that would distinguish clear air turbulence and trailing vortices from other sounds.
“How nature works is that each event has a unique signature. Clear air turbulence has a unique signature; the wake of each aircraft has a unique signature,” Shams explains. For in-flight turbulence detection, an algorithm might take a sound detected by the plane’s mic array and compare it to the library of signatures, looking for a match or close match that would clearly identify that sound as trailing vortices and not, for instance, a distant rocket launch.
To build the library, they needed to capture sound signatures of trailing vortices from passenger and military aircraft flying in and out of a local airport. In 2013, Shams received permission to dig three holes with shovels about 80 meters from a runway at the Newport News-Williamsburg International Airport in Virginia to install microphones in boxes below ground. Another set of microphones was installed in portable spherical containers to create an above-ground array about 150 meters from the runway. Trailing vortices signatures were collected for about two years. This chapter was exciting, but also tragic: Zuckerwar died in 2014 after a long battle with cancer.
To push the research forward, Shams knew he needed to capture sound signatures with microphones installed on an aircraft. He and colleagues decided to start with trailing vortices. This is when Shams decided to team up with Brown, the research pilot at the Aerospace Research Centre in Canada. The two met in 2019 not long after Brown saw a presentation by Shams about his airport mic installation. Brown had been tracking wake vortices, contrails and emissions under Canadian research projects by flying a Lockheed T-33 behind airliners. Agreement was reached to install one of Sham’s mics on the T-33 to collect trailing vortices signatures.
Why start with trailing vortices? Because, unlike clear air turbulence, it was obvious where to find them — behind airliners — and the acoustic characteristics could be readily recorded.
“You look for strong sources where you can identify the vortices and therefore have confidence in the microphone response from an identified source,” Brown explains.
They needed to install the mic with as little modification to the T-33 research plane as possible, so they connected it to a tube fed from one of the plane’s pitot tubes, the small, L-shaped appendages on fuselages that measure air pressure and wind speed. Voltage from the mic was wired into an onboard computer to collect the infrasound. Because the mic was extremely sensitive to the wind noise in the pitot tube, they capped the mic with a thin layer of closed-cell foam. One mic was sufficient for their flight tests, but the future concept for a commercial airliner would probably require two mics on the wingtips and one on the tail to triangulate the direction from which the sound was coming, Shams says.
Brown, who grew up in Tasmania and still speaks with an Australian accent and drops “mate” into his conversations, has flown commercially and also for the Royal Australian Air Force. To record infrasound from trailing vortices, Brown picked his targets with the help of air traffic controllers for the Ottawa, Ontario, airport. They notified him of the flight paths of large passenger jets flying between Europe and Toronto, and Europe and U.S. cities on the West Coast or in the Midwest. Before each of his two-hour flights in the T-33, he picked out five airliners that would be passing through at an altitude of 35,000 to 38,000 feet and aimed to track three of them. He climbed to their altitude and picked one to approach from the front, recording its infrasound signature while he flew within 8 kilometers, passed 2,000 feet below the airliner and then trailed it at a distance of 8 kilometers—the separation distances required by flight rules.
As Brown trailed, flying in the teeth of the wake turbulence, he also recorded instantaneous wind speeds and vectors within the vortices at a rate of 600 times per second, employing a data collection system he had designed for the T-33. Combined with the infrasonic signatures, the detailed wind data created a more complete picture of the turbulence and its potentially dangerous forces behind a large airliner—forces powerful enough to flip a smaller aircraft or even extinguish jet engines.
“They’re very susceptible to wake vortex, blowing them out like a candle,” Brown says. “You get a real rough ride. Wake vortices are particularly nasty.”
The T-33 flights ended in October 2021 after collecting trailing vortices data from about 30 airliners of various makes.
During the project, the company Stratodynamics of Delaware licensed the detection technology from NASA and developed an infrasonic technology it calls Vortesight. The company is now seeking investors to help develop it into a commercial turbulence detection product, says founder and CEO Gary Pundsack. He estimates the technology will need another two years more of development before it’s ready for market, and then more time to win FAA approval.
For the NASA-Canada collaboration, the next step will be difficult: recording infrasonic signatures of clear air turbulence and measuring the size of these turbulent areas and the strength of their air currents. The hard part will be finding instances of clear air turbulence because they can’t be predicted through weather forecasts. Brown expects jet streams flowing at altitudes between 3,000 and 6,000 feet are most likely to produce the required conditions.
Rather than hunting for clear air turbulence, Brown and his colleagues plan to capitalize on flights they began in December with a De Havilland Twin Otter turboprop for other kinds of research. They’ve added an infrasonic mic to the plane’s instruments so that when clear air turbulence is eventually encountered, they will have a record of its sound.
Collecting the sounds of clear air turbulence will take longer than doing so for trailing vortices data, which took less than a year, Brown says. The collaborative research agreement between NASA and NRC is scheduled to last through 2027.
“The Twin Otter will hopefully — not all the time, but sooner than later — fly into some good meteorological shear fields that have good vortex structures,” Brown says.