The Doppler road show hits it big at
This is the second year for a major field program studying tornadoes in the U.S. southern plains. The storms were there in force, and some powerful new equipment for studying them produced exciting results. In its first test with all its capabilities functioning, NCAR's new airborne doppler radar performed well beyond expectations. On the ground, a mobile radar called Doppler on Wheels played tag with tornadoes and caught detailed images of their inner workings. (See center spread for more photos and images from both radars.)
When meteorologist Roger Wakimoto first heard about the new high- resolution airborne Doppler radar that Peter Hildebrand and colleagues in NCAR's Atmospheric Technology Division were building, he called it "tailor- made for studying severe thunderstorms."
Now that idea has become reality. Under Hildebrand's guidance, ATD's Research Support Facility constructed the Electra Doppler radar while the NCAR/NSF Electra aircraft underwent extensive modifications to carry it. There were refinements and a deployment of the not-quite-finished ELDORA in a major field project in the western Pacific in 1993--followed by more refinements. This spring was its first deployment at full capacity. With Wakimoto on board, the Electra took ELDORA to the brink of tornadic supercells on the southern plains. The radar performed "at a marvelous level, high above expectations," says the meteorology professor from the University of California, Los Angeles. "This is as good as news gets from the instrument- development point of view."
ELDORA's coup is only one of the success stories emerging from VORTEX, the Verification of the Origin of Rotation in Tornadoes Experiment. This intensive study of tornadic and nontornadic supercells completed the second of two spring field phases on 15 June. Over 100 people from eight universities and several laboratories assembled at NOAA's National Severe Storms Laboratory (NSSL) in Norman, Oklahoma, for the experiment. Other key players besides NSSL were the University of Oklahoma (OU), its Center for the Analysis and Prediction of Storms (CAPS), the National Weather Service, and UOP's Office of Field Project Support (OFPS). Aircraft included the Electra, a NOAA P-3 with dual Doppler radar, and the T-28 from the South Dakota School of Mines and Technology. Besides ELDORA, there were two new ground-based portable Doppler radars. One, owned by the University of Massachusetts, was operated by OU's Howard Bluestein. The other was Doppler on Wheels, constructed by OU assistant professor Joshua Wurman with support from OU, NSSL, the NSF, and NCAR. The principal investigators included director Erik Rasmussen (NSSL) and assistant director Jerry Straka (OU).
VORTEX approached a basic question--what makes some thunderstorms produce tornadoes and others not?--with a set of finely tuned hypotheses and an observation network gathering what Rasmussen calls "large, contextual, comprehensive data sets." The linchpin of the observing network was a set of 21 vehicles, 16 equipped with roof-mounted instrument packages designed by Straka, Rasmussen, and NSSL's Sherman Fredrickson. The VORTEX armada, staffed by students and visiting scientists, made a dramatic impression as it rolled along highways and back roads through the plains. (On one April afternoon last year, a nervous official in a small Texas town asked the armada to move--their presence was making the residents jumpy.)
Data appeared every five seconds on laptops in the vehicles, whether at rest or in motion. The field teams relayed locations and conditions via telemetry and very high frequency government radio channels to a mobile coordination center that directed the movement of cars and planes. Software designed by Rasmussen for the command post displayed all the teams' locations in real time along with road maps and projections of storm motion.
Rendezvous at Garden City
ELDORA had been built at NCAR, with the French government contributing a
radome, antenna, signal processor, and some design work. By early 1993, it
was ready for deployment in the Tropical Ocean and Global Atmosphere
program's Coupled Ocean-Atmosphere Reponse Experiment (TOGA
COARE), albeit with only one of its two transmitters working. (Even so, the
system made a useful contribution to the program.) In the following two years
NCAR completed a major overhaul of some troublesome hardware and added
the radar's most advanced features. "We can now produce the high-
resolution dual Doppler data that ELDORA was designed to collect," says
The first big test for the new ELDORA came on 16 May, which was also the first day that a newly installed weather avoidance radar data system was in place. The system made use of compact signal processing created by NCAR's Mitchell Randall that allowed for weather data to be gleaned from the navigational radar in the Electra's nose cone (see UCAR Quarterly, Spring 1995). The two radar systems were an unbeatable combination. After a few passes around a promising storm near Garden City, Kansas, recalls Hildebrand, "the nose radar immediately indicated areas of circulation in the storm that evolved into hook echoes over the next half-hour or so." Hook echoes are hook-shaped patterns in radar images that can foreshadow or accompany tornado formation. Throughout the flight the researchers used the nose radar to seek out areas where tornadoes might form.
ELDORA and the NOAA P-3 radar ended up capturing 90 minutes of data prior to formation of a relatively weak tornado, estimated at F1 to F2 on the Fujita scale (winds between 33 and 70 meters per second). The tornado stayed on the ground for another 10 to 15 minutes, tracing a path about seven kilometers long. "With a fixed Doppler radar, you may only collect about 20 minutes of good data at best," says Wakimoto. "We were able to follow this storm across a good fraction of the state." According to Rasmussen, the coordinated efforts of the Electra and P-3 on four tornadoes observed this spring should provide "data of much higher quality than either aircraft could provide alone."
Wakimoto notes that he and other members of the flight team were working with ELDORA "on the fly" to estimate parameters such as the strongest winds that could be measured unambiguously. They settled on 79 m/s, far higher than the usual 10 to 15 m/s. Thanks to the ELDORA's rapid scanning and fine resolution, the result was "a very clean data set," says Wakimoto. He adds that the shakedown period "has cost us a little bit of data. At times we've chosen the wrong parameters, flown at the wrong height. Of course, none of us had any prior guidance. This year was the precedent-setter."
Immensely helpful in the learning process, says Wakimoto, was the analysis room housed at OU that served ELDORA as well as Doppler on Wheels. "We knew if we didn't evaluate data right after a flight, we'd be in big trouble. For the first time in my experience, we had a real data center where scientists could sit down, review data, and discuss the next flight. It's exactly the way a joint UCAR-university-scientist center should be."
By the end of VORTEX, the Electra had used most of its 96 allocated flight hours. "I wish there had been one or two more tornado cases," says Wakimoto, "but I feel quite good about the project. It takes a lot for me to be impressed, and I'm very impressed with ELDORA."
Doppler on Wheels
Doppler on Wheels begins a new generation of mobile Doppler units deployed in Tornado Alley. OU's Bluestein pioneered the approach, measuring 128-m/s winds in a tornado on 26 April 1991 with a continuous-wave unit that transmits constantly rather than in pulses to give a generalized picture of velocities in an area. Wurman, Straka, Rasmussen, Randall, and Alan Zahrai (NOAA Operational Support Facility) scavenged equipment from NSSL and NCAR to extend the mobile strategy. Their creation--built at NSSL in less than four months--is more akin to standard research radar than previous portable Dopplers. The use of X-band frequencies (around three centimers in wavelength) allows for penetration of heavy rain; a pulsed, scanning signal and large-dish antenna permit resolutions of a few dozen meters. The radar's portability is enhanced by Randall's technique for putting formerly bulky signal-processing tasks into a standard personal computer.
"To assess tornadogenesis," says Rasmussen, "one needs to scan the entire low-level mesocyclone region [storm-scale and smaller circulations that can spawn tornadoes] every 90 seconds or less. We were able to do this in VORTEX."
A team including Wurman, Straka, and Randall brought Doppler on Wheels within a few kilometers of a sunset tornado near Jetmore, Kansas, on 16 May. But a violent twister arcing around Dimmitt, Texas, on 2 June was the prize catch.
Climax at Dimmitt
The Doppler on Wheels crew parked about three kilometers north of the slow-moving, long-lived tornado, which traversed an east-to-north arc during the 11-minute observation period. A damage survey showed the storm's violent character: one 40-meter section of asphalt roadway had been torn away, along with power lines and a few structures. Nearby were what Wurman called "pretzelized cars." (Fortunately, the tornado bypassed most of Dimmitt itself.)
The collaborators were more than satisfied and intrigued by what Doppler on Wheels found. "It actually looks like a hurricane," notes Wurman. "There's an almost-circular eyewall and an almost-vertical eye, absent in debris near the ground and 300 meters in diameter near the cloud base." Wind contrasts of more than 120 m/s across the circulation were observed, along with asymmetries such as a stream of inflowing air similar to those seen in hurricanes. Though full analysis will come later, Wurman says "the data seem to agree with some theoretical predictions" on tornado circulation. Though the Electra and P-3 were somewhat farther from the Dimmitt storm than usual, they did obtain data, providing multi-Doppler coverage.
Putting It Together
Much of the VORTEX data as a whole has already been archived by OFPS.
The UCAR-based office has become a specialist in the processing and
cataloging of data from giant field projects, including TOGA COARE and the
Stormscale Operational Meteorology Program's Fronts Experiment Systems
The main tool by which OFPS helps scientists to access data is now available on the World Wide Web. Called the Codiac data management system, it was devised three years ago and converted more recently to a Web interface. Over the past year OFPS oversaw the deployment of soundings during the two VORTEX field phases while putting a catalog of the experiment's data on line through Codiac. The office is working with NSSL on quality control for the final data set.
Part of that set will be output from a mesoscale model devised by CAPS. Its Advanced Regional Prediction System was used for the first time in a real- time fashion during VORTEX, updating twice-daily NWS model output with more-frequent surface and upper-air reports. Forecasters in the Storm Prediction Center at NSSL guided VORTEX operations as they experimented with new forecast techniques, making use of the temporary wealth of data spawned by the experiment.
One of the biggest puzzles to be sorted out in the post-field phase is why so few of the major storms observed by VORTEX produced tornadoes. A long- standing estimate was that one out of every two southern plains supercells produced tornadoes. The estimate is now much smaller. Two mammoth supercells, one in northern Kansas on 11 May and the other near Shamrock, Texas, on 22 May, particularly intrigued Wakimoto. "Both of these were superb visually, rotating like crazy, but did not produce tornadoes over periods of two or three hours."
Tornadic or not, the power of these two cells was not lost on the ground crews. Hail as large as softballs pummelled the vehicles on both days, destroying several windshields. Even people on the Electra, which stayed just beyond the perimeter of the storms it studied, sat up and took notice. "These were some of the roughest flights I have ever been on," says Hildebrand. "While the biggest bumps didn't seem at all aircraft-threatening, we flew through strong turbulence and made tight turns for five to seven hours a flight. Most passengers ended up having an intimate relationship with their air sickness bags."
See the VORTE X home page,, on the World Wide Web, which includeds of Doppler on Wheels tornado imagery and the OFPS catalog of data compiled to date.
ELDORA captured high-resolution images of a monster tornado more than a kilometer wide that struck Friona, Texas, on 2 June. At the time these images were obtained, the tornado was about 25 kilometers from the aircraft (the concentric rings are at 10-kilometer intervals). At top, the reflectivity image clearly shows the location of the tornado, characterized by the weak echo tube extending from the surface up to 14 kilometers. Radar images of tornadic storms usually show a core of high reflectivity (the signals returned to the radar by intense precipitation, such as large raindrops and hail) but are unable to resolve the weak echo inside a tornado. The center image, Doppler velocity, reveals the velocity structure of this tornadic storm. The color scale is centered at 37 meters per second to correct for the ground speed of the aircraft. Therefore, warm colors (reds, oranges, and yellows) indicate winds moving away from the plane; cooler colors (blues and greens) are winds blowing toward it. The bottom image is spectrum width, an indicator of turbulence and shear within the sampled volume. Note the high spectrum width (around 12-15 m/s) associated with the tornado signature. "What we gained in this experiment," says ATD researcher Wen-Chau Lee, "were consecutive scans 300 meters apart for the first time. There is also unambiguous resolution of velocities in the range of plus or minus 79 meters per second, compared to the previous standard of plus or minus 13 meters per second." This means that ELDORA can resolve a much wider range of wind velocities. (Image courtesy of Roger Wakimoto and the ATD Remote Sensing Facility.)
Only days after deployment, these striking images of preliminary data gathered by Doppler on Wheels were on the World Wide Web.
Velocity depiction from a storm near Jetmore, Kansas, on 16 May, 8:43 p.m. local time. The tornado is at the center of the velocity couplet (region with strong and contrasting winds side-by-side), around ten kilometers northwest of the radar, whose elevation angle is at one degree.
Reflectivity depiction from the storm near Dimmitt, Texas, on 2 June, 8:08 p.m. local time. The tornado--more than a kilometer wide--is centered near the 3-km range ring toward the top of the image. The tornado's outer eyewall is indicated by the yellow ring, while an inner core of lower reflectivity (less precipitation cloud material, or debris) appears bright blue. This image was taken at 18 degrees elevation, or about a kilometer above ground level at the tornado. (Illustrations courtesy Joshua Wurman, Erik Rasmussen, and Jerry Straka.)
Doppler on Wheels is deployed near Dimmitt, Texas, with a violent tornado (left) only two kilometers distant. (Photo courtesy VORTEX.)