UCAR > Communications > Highlights > 1998 Search


A Turbulent Situation

Your flight attendant tells you to fasten your seat belt.
Ten seconds later, your coffee's on the ceiling.

The stealthiest of weather hazards, turbulence can strike from a sky that is literally clear and blue. It has brought planes to the ground and stymied some of the greatest minds in physics. At NCAR, a long-term program in turbulence theory continues to nourish new approaches to the problem, while an applied project is using passenger planes as turbulence sensors to create the nation's first quantitative map of hazard zones. Bridging the realms of theory and reality are computer simulations, revealing sides of turbulence that scientists didn't know existed.


In drawings, Leonardo da Vinci depicts turbulent vortices.


Osborne Reynolds, a British scientist, develops the concept of transition from viscous to turbulent flow, using a single parameter--the Reynolds number.


German scientist Z.A. Prandtl identifies mixing-length theory, the methodology still used today to identify the scales at which turbulence is important relative to other forces.


Britain's L.F. Richardson develops the concept of eddy breakdown across a range of scales, a concept he relates in both equations and verse:

Big whirls have little whirls,
That feed on their velocity;
And little whirls have lesser whirls,
And so on to viscosity.


The Five-Thirds (5/3) Law, relating the scale of turbulent eddies to the amount of energy within the eddies, is derived by Russian scientist Andrei Kolmogorov.


Edward Lorenz (Massachusetts Institute of Technology), a frequent NCAR visitor, sets limits on the predictability of weather systems by developing chaos theory. In a 1972 paper, Lorenz pondered whether a butterfly flapping its wings in Brazil could trigger a tornado in Texas.


Douglas Lilly and James Deardorff (NCAR) and Joseph Smagorinsky (NOAA) develop large-eddy simulation (LES), the first practical technique that accounts for turbulence in numerical models.


In simulations at NCAR, Stuart Patterson and Cornell University's Eric Siggia and Robert Kerr (then a visiting graduate student, now an NCAR scientist) identify vortex tubes within turbulent flow.


NCAR launches its Geophysical Turbulence Program.


United Airlines begins testing a turbulence sensing system devised at NCAR.

It was a windy morning across the Front Range of Colorado--not uncommon for late fall. The high-altitude jet stream and an accompanying cold front were moving across the Rockies from the west-northwest. Cruising just west of Denver in that river of air was a coast-to-coast DC-8 cargo jet. It crossed the Front Range about four miles above the highest peaks, normally high enough to avoid mountain-induced turbulence.

At about 9:00 a.m., without warning, severe updrafts and downdrafts buffeted the DC-8 for two minutes. One engine and part of the left wing were sheared off. The aircraft hobbled into Stapleton International Airport for an emergency landing.

The plane's pilot was lucky to have survived. The event was a lucky one for research meteorology, too, because on that morning--December 9, 1992--an exceptionally wide array of instruments happened to be trained on the skies of Colorado. A high-altitude layer of dust, left over from the 1991 eruption of Mt. Pinatubo, helped scientists from the National Oceanic and Atmospheric Administration (NOAA) to track winds from the ground using a lidar (laser-based radar) stationed near Boulder. It was a routine night of sampling that produced results that are anything but routine.

"The pilot didn't expect to hit what he hit," says NCAR physicist Robert Kerr. A satellite photo taken only two minutes after the accident provided few clues to what had happened along the flight path. At lower levels, "There was a huge rotor," says Kerr, "but the pilot knew how to avoid that." Rotors--lengthwise circulation tubes formed as flow ascends and descends mountains--are a common feature during windstorms, often made visible by lenticular wave clouds.

The mystery encouraged Kerr and NCAR computer scientist William Hall to feed reams of data from NOAA and other sources into a small-scale computer model of atmospheric dynamics created by Hall and NCAR senior scientist Terry Clark. Processing the job took several days on one of NCAR's CRAY J-9 parallel supercomputers, followed by additional processing at the center's Scientific Visualization Laboratory. When the results came back, the team was astonished to find tubes of circulation stretching eastward at high altitudes from the mountains, near flight level. Were the vortex tubes real, or figments of the model's imagination?

A closer look at the satellite photos revealed cloud streaks at roughly the same level as the aircraft incident, where the model had indicated strong vorticity, or rotation. When the model's output was pushed to its highest resolution--200 meters (660 feet) between points over a
50-km (31-mi) domain--it showed enormous wind shear on the edges of vortex tubes. From one point to the next, the wind speed changed by as much as 40 meters per second (90 mph).

Graphic courtesy Don Middleton, NCAR Scientific Visualization Laboratory
Areas of strong jet-stream winds (blue) and rotating wind shear (yellow) are highlighted in a computer model's depiction of conditions at the time of an aircraft incident above Colorado's Front Range. The purple area at center is a vortex tube of the sort being studied by turbulence modelers at NCAR. This graphic was produced at NCAR's Scientific Visualization Laboratory. With an array of cutting-edge tools for putting complex data into three-dimensional imagery, the "Viz Lab" is a powerful tool of NCAR and university scientists.

Given that the actual vortex was probably smaller than the model's lowest resolution, "We have every reason to believe the [circulation] could have been five times stronger than that," says Kerr. If so, it was as tightly packed and powerful as a modest tornado. But these vortex tubes, and other turbulent phenomena like them, occur outside of rain or snow, with nothing visible to warn pilots of their presence.

Partly because it hides so well, turbulence remains an enigma to atmospheric science. The famed physicist Richard Feynmann called it the last great unsolved problem in classical physics. "Feynmann said that in the 1950s, and it's still true today," Kerr says. NCAR's founders recognized that turbulence was central to a better understanding of the atmosphere. Throughout the 1960s and 1970s, a Turbulence Club met regularly to hash out approaches to the problem. Former NCAR scientists Doug Lilly (University of Oklahoma) and James Deardorff (Oregon State University) came up with a way to depict turbulent eddies in a computer model. Instead of trying to model every swirl down to the smallest scales, one could describe statistically how the effects of a group of small eddies might affect larger-scale motion. This technique, called large-eddy simulation (LES), is a cornerstone of turbulence modeling to this day.

LES relies on statistics to describe turbulent behavior, but thanks to ever-more-powerful computers, it's now possible for some simulations (like the one created by Clark, Hall, and Kerr) to depict turbulence more directly. Until now, colorful names--pancakes, worms, spaghetti, and noodles--have denoted structures found only in mathematical theory or in idealized simulations. These phenomena are "very intermittent, and they occur at much smaller scales in the atmosphere than we could represent on our computers," says NCAR senior scientist Jackson Herring.

Herring arrived at NCAR in 1972, full of hope. "At that time, I think we were more optimistic than now that we'd find a general theory of turbulence," he recalls. Despite the lack of an overarching theory, approaches such as LES have proven useful, especially in studies of the planetary boundary layer (the atmosphere's lowest kilometer) and other regions where turbulence is a key player.

Given that turbulence research is, in Herring's words, "a difficult field with different people doing different things," the need for collaboration and communication is great. Small but vital, NCAR's Geophysical Turbulence Program (GTP)--a formalized version of the older Turbulence Club--brings scientists from virtually every part of the center together with colleagues from around the world. Members of the program hold a monthly seminar, sponsor one or two scientific visitors each year, and hold a major meeting every year or two. One of the largest occurred in June 1998. Cosponsored by GTP and two international scientific unions, it brought over 100 of the world's turbulence experts to Boulder for four days of discussions and presentations.

William Hall (left) and Robert Kerr used observations and computer models to explore the role of turbulence in a 1992 aircraft incident above Colorado's Front Range.

It's possible that some of the attendees at this symposium flew into Denver on aircraft that, unbeknownst to them, served as turbulence sensors. Starting in 1997, a few United Airlines passenger planes were fitted with software that converted every bump and jostle of the aircraft into a numeric reading of turbulence. More than 600 United, American, and Northwest aircraft may have these turbulence sensors by 2001. It's the first time airlines and scientists have had a quantitative, comprehensive picture of this persistent hazard.

The plane-based sensors are the brainchild of NCAR scientist Larry Cornman. With funding from the Federal Aviation Administration, Cornman and colleagues have identified a way to make use of the plane's existing sensors, computers, and communications systems, thus avoiding a costly retrofit. The aircraft's response to turbulence is sensed and the turbulence itself inferred through a technique that accounts for the plane's vertical acceleration, weight, air speed, altitude, and autopilot status. Data are taken once every minute and sent directly to NCAR. By late 1998, the center will be producing a Web-based graphics product and sending it to airlines and NOAA's Aviation Weather Center. A map will outline flight paths across the nation and display regions where turbulence is reported.

The new system marks a vast improvement over standard qualitative pilot reports that note only "light," "moderate," or "severe" turbulence. "While this information is somewhat useful in an aviation safety context, it is of very little use to the atmospheric science community," says Cornman. In June 1998, the new measurement technique got a vote of global confidence: the International Civil Aviation Organization voted to make it the international standard for reporting turbulence from commercial aircraft. After approval by member nations, the action will become official by 2001.

For all its strengths, the new measurement system only reports turbulence in progress--it doesn't see it coming down the pike. Another milestone in turbulence sensing was reached in April 1998, as an NSF/NCAR aircraft probed the bumpy skies above Denver for the National Aeronautics and Space Administration (NASA). The C-130 carried a lidar (built by Coherent Technologies, Inc.), that successfully tracked air motions--and thus turbulence--up to several kilometers ahead. Conventional radars already used aboard aircraft to detect wind shear may also prove useful in detecting some kinds of turbulence, says Cornman: "Lidar is good for finding clear-air turbulence, but not in-cloud turbulence, while radar is just the opposite."

Placing lidars on board the nation's commercial aircraft would be a major undertaking, but Cornman foresees a day when the two new sensing systems could be combined to forge a complete detection-and-warning system. NASA has embarked upon an aviation safety program that aims to reduce turbulence incidents fivefold within a decade. Cornman's group at NCAR is already developing a prototype display that will show pilots in the cockpit the locations along their routes where they might encounter turbulence, icing, and other currently hard-to-map hazards. Meanwhile, colleagues are working on new methods for turbulence prediction based on the output from numerical computer models for weather forecasting.

NCAR's aviation work shows that, even without a complete understanding of turbulence, progress is possible. Several laboratories have more extensive programs in turbulence theory, but none is as oriented toward the atmosphere as NCAR's. And, as Cornman sees it, "In some senses I'd say we're the world's premier center for applied turbulence work, especially with respect to aviation. We were in the right place at the right time, doing good work. People come to us now."

On the Web

NCAR/Geophysical Turbulence Program
NCAR/Aviation Weather Development Laboratory

HL Contents The Eleven-Year Switch Particles of Doubt A World of Cycles A More Perfect Science The Art of Counting Raindrops A Turbulent Situation Wired for Weather

Other issues of UCAR Highlights


Edited by Bob Henson, bhenson@ucar.edu

Prepared for the Web by Jacque Marshall
Last revised: Mon Apr 10 13:23:27 MDT 2000