Langley researchers try to observe the turbulence radar during a bumpy encounter near thunderstorms on board the ARIES 757 research aircraft.

	NASA’s Airborne Research Integrated Experiments System flew from Langley east of the Mississippi River to find atmospheric turbulence. The 757 made 13 bad-weather flights in spring 2002 to test software/hardware upgrades to existing weather radar that allow it to detect turbulence associated with storms.

Stormy weather for testing
NASA engineers and software designers tested the turbulence detection system over a series of 13 flights last April and May aboard the agency’s Airborne Research Integrated Experiments System, a Boeing 757 airliner equipped for avionics research.

On each mission at Langley, the plane circled thunderstorms while researchers at test stations recorded atmospheric conditions. They told the pilot when the enhanced software predicted turbulence, and how strong it would be.

Over the coming months, NASA software designers will refine the enhanced radar processing technique and the automated warning system before turning over the nuts and bolts of the software to commercial avionics companies. It will be up to the airlines and their contractors to decide whether to add the new capabilities to existing aircraft weather software.

“We’ve gotten enough information from these flights that between now and ’05, when the program is supposed to wind down, we can do another flight—in late ’04 or early ’05—and put together all the lessons we’ve learned,” says Rod Bogue, manager for the turbulence prediction and warning system program at NASA-Dryden.

NASA engineers do not expect the FAA to order carriers to upgrade their aircraft with the turbulence software enhancements, as regulators did with the wind-shear prediction radar. They do hope, however, that airline executives will see the wisdom of installing the software, which could be ready by 2004 or 2005, says O’Connor.

Background
Engineers from Langley and Dryden began working on the turbulence detection and warning software in 2000. The program sprang from a $500-million effort by NASA and the FAA to improve aviation safety.

The agencies undertook the Aviation Safety Program on the recommendation of the 1997 White House Commission on Aviation Safety and Security, led by then Vice President Al Gore. The Gore commission called for cutting the rate of aviation accidents by a factor of five over 10 years.

The turbulence effort falls under the Weather Accident Prevention Project, which is part of the Aviation Safety Program. Winning funding for such work was difficult, because turbulence is rarely fatal to airline passengers or crews. It “puts a lot of people into emergency rooms, but not into morgues,” says Bogue.

Three people have died in turbulence-related accidents over the last 25 years, but many more were injured, says Bogue. Indeed, turbulence is the number-one cause of in-flight injuries to passengers and crews. Each year, it injures 300-350 people seriously enough to warrant trips to the hospital, notes Bogue. Ninety-eight percent of the victims were not wearing seat belts. Flight attendants are most susceptible, because they are often the last to sit down and buckle up when a plane encounters rough skies, he points out.

The goal of the program is to offer pilots up to 2 min of warning time so they can direct passengers and crew to return to their seats and buckle up, Bogue says. So far, NASA engineers feel confident they can provide up to a minute of warning with the enhanced software.

	The enhanced radar system shows turbulence to pilots in a graphical display.

Getting an earlier alert
Hidden behind the nose cone of every U.S. airliner is a radar antenna that scans the sky for moisture out to a range of 16-20 mi., says Bogue. Severe storms are depicted as shades of red on cockpit weather displays. Pilots can avoid some turbulence by steering clear of these storms. However, deciding when to warn passengers about the problem is mostly a matter of intuition. The only turbulence warnings pilots currently receive are informal radio reports of “chop” or “bumps” reported by pilots flying ahead of them. Pilots overhear these reports on their radios, or sometimes the alerts are relayed through flight centers, say NASA engineers.

The software that processes the data from an airliner’s radar is not sensitive enough to detect moisture in the relatively dry air that rises and falls many miles from storms or swirls over mountain peaks.

When buffeting has no obvious source, pilots call it “clear-air turbulence.” The term is something of a misnomer, however, because most turbulence is still caused by convection—the process of warm air roiling its way through the atmosphere. “True clear-air turbulence is a rare occurrence. Just about all the turbulence aircraft encounter is convectively induced,” says Jim Watson, the deputy turbulence prediction and warning systems project manager at Langley. “The moisture is so low you just don’t see it.”

In some cases, turbulence-producing storms are not visible at all on weather displays, because the aircraft’s radar scans straight ahead at the flight altitude. “There can be a thunderstorm building up that you don’t see. You fly over it and you get hammered,” Bogue says.

The new software consists of mathematical algorithms that make the existing weather radar more sensitive to low levels of moisture. By tracing the movements of moisture in air that otherwise appears clear, the radar can detect turbulence that was hitherto invisible, NASA engineers say.

The enhanced software would give pilots their first visual detection of turbulence and predict its severity. If a plane with the software entered turbulence, its computer also would convert readings from onboard accelerometers into actual “in-situ” turbulence data. The computer would rank the severity of the jostling and transmit automated warnings to planes following behind.

The computer on a receiving plane would scale the strength of the turbulence to the size of the plane. “You don’t want to alert a commercial pilot with a report from a Cessna. But if you’re in a Cessna, you want to know if a 747 just got rocked,” says O’Connor, who managed development of this in-situ warning system.

Improving on Doppler radar
When they undertook the program, NASA engineers knew cash-strapped airlines would not spend a lot of money upgrading aircraft with new hardware to detect turbulence. They decided to capitalize on the computational and radar upgrades ordered by the FAA in 1988 after a spate of airline crashes that were blamed on wind shear. A particularly dangerous kind, called a downburst, happens when cool air rushes downward out of severe storms and collides with the ground. When aircraft are descending toward the runway, they are especially vulnerable to these downbursts, Bogue says.

Commercial airliners are now equipped with Doppler radar software that predicts low-altitude wind shear by looking for telltale upward spikes in the radar frequency, or Doppler shift, Bogue says. Wind shear is similar to turbulence in that both occur in the relatively clear air surrounding storms, he adds. Both require radar processing that is sensitive to tiny levels of moisture.

“The challenge is to sift that Doppler increase out of the low signal-to-noise ratio,” Bogue points out.

The turbulence avoidance system would amount to an enhancement of the existing wind-shear predictive Doppler radar system mandated by the FAA but currently used only near the ground.

The engineers tested the enhancements over a series of Boeing 757 flights last spring, with encouraging results. “We feel we can predict turbulence in front of an aircraft in the 30-60-sec range,” Watson says.

During one encounter, the enhanced software predicted severe turbulence in the clear air ahead; 50 sec later, the 757 dropped. “We fell about 1,000 ft, from 30,000 ft. Falling 1,000 ft in a huge commercial airliner is not desirable. It was difficult for the crew to even hold a video camera, and the pilot said it was difficult to read the gauges,” Watson explains.

Back at Langley, the engineers compared the actual severity of the turbulence, as measured by the plane’s accelerometers, to their predictions. “We got very high correlation,” Watson adds.

Bogue points out that an additional flight is needed to test refinements that engineers are currently making to the software. He says the ultimate goal is to provide up to 2 min of warning.

O’Connor also tested the in-situ warning system during the flights by transmitting actual turbulence readings to a satellite, then to a ground station, and then back to the 757, as if it were a different plane. “It took 1.5 sec to make the whole loop. You can’t beat that,” O’Connor says. O’Connor hopes to conduct a more elaborate experiment in which a smaller jet would follow behind NASA’s 757 research plane. “The 757 would go through and a Lear jet [-class plane] would follow at a safe distance of 5-7 mi. Commercial jets use these conga lines all the time,” he says.

That test would demonstrate the software’s ability to scale the warning to a smaller plane. “The computer would convert [the turbulence] to the right scale and interpret whether there should be a warning or not,” O’Connor says.

Other approaches
The engineers also are studying other, far-reaching technologies that show promise for detecting turbulence. One option would be to use a light detecting and ranging (lidar) device. Lidar, which is more sensitive than radar, bounces light waves rather than radio waves through the atmosphere. It is used by atmospheric scientists to detect aerosols, tiny dust particles invisible to radar.

For radar to spot turbulence, at least some moisture is required, Bogue notes. “Sixty percent of turbulence is convective turbulence. In those cases, you have enough moisture in the atmosphere to detect the turbulence at a reasonable rate with the [enhanced] radar approach. “But what do you do when you don’t have any moisture?” Bogue says. “One idea is to use lidar.”

Issues with lidar are the volume, power, and weight of the equipment needed to generate and focus the light. “We’re still looking at power and range with lidar. But it’s fair to say there’s quite a bit of potential,” he continues.

Even so, Watson says it makes sense to focus on radar at the moment, because it uses existing airliner hardware and computational muscle. “The weight and the power requirements and the fuselage modifications that it takes to install one of these [lidar] systems are outweighing any of the benefits we’ve been able to identify,” Watson says.

NASA engineers have just started looking at another possible option that would use airliner lightning detectors to perceive turbulence. When planes are flying in clear air, the lightning detector sometimes senses an electrical discharge short of lightning, Bogue says. “There seems to be a connection between unstable electrical discharge in clear air and turbulence,” he notes.

Although this phenomenon is worth exploring, it is too early to tell whether it could lead to another low-cost way to detect turbulence, Bogue points out. “There are more things we don’t know about it than we do know,” he says.