A new spin on tornado science

Researchers take to the plains to unlock secrets of twister formation

April 7, 2009 | When the movie Twister became a global blockbuster in 1996, researchers had just completed two seasons of real-life chasing in the world’s largest ever tornado study. The Verification of Rotation in Tornadoes Experiment (VORTEX) saw scientists and technicians from NCAR and elsewhere roaming the Great Plains in 1994 and 1995, documenting the conditions that lead to tornado formation (tornadogenesis).

Thunderstorm east of Plainville, Kansas ESSL/MMM visiting scientist Howie Bluestein (University of Oklahoma) took this photo east of Plainville, Kansas, on May 29, 2008, while tracking the thunderstorm at left with a University of Massachusetts radar (center). The storm went on to produce several tornadoes.

There was no sequel to Twister, but a second VORTEX study is unfolding this spring and next. Funded mainly by NSF and NOAA, VORTEX2 will bring several times more mobile platforms to the plains, along with updated research questions that build on the findings of the first VORTEX.

“We still do not completely understand tornadogenesis,” says EOL director Roger Wakimoto. Roger was a principal investigator on the original project while a professor at the University of California, Los Angeles. “It’s our hope that VORTEX2 will finally provide some of the answers we’ve been seeking,” he adds.

Doppler radars installed at U.S. airports through the 1990s have produced a revolution in tornado tracking. The average lead time for warnings more than doubled from 1993 to 2005, increasing to 13 minutes, according to the National Weather Service. Yet these warnings are far from perfect, with false alarms still a problem. The strongest tornadoes typically leave an obvious signature on Doppler radar, but in many other cases the signals are ambiguous. Thus, forecasters must often issue warnings based on a suspicious looking radar feature without knowing for sure whether a tornado is actually occurring.

David Dowell

David Dowell.

 “Without very high resolution observations, it’s very hard to tell the difference on radar between a supercell thunderstorm that produces a tornado and one that doesn’t,” says David Dowell (ESSL/MMM and RAL). In the original VORTEX study, scientists had hoped that the wealth of wind data collected from car mounted weather stations, radiosondes, and the Electra Doppler Radar (ELDORA) would clarify the difference. But the crystal ball remained fuzzy.

Poring over the VORTEX data, investigators concluded that strong contrasts in temperature and moisture—spanning distances of a few miles or less and unfolding in a matter of minutes or even seconds—could be the factor that allows one storm to produce tornadoes while a similar looking storm doesn’t. To get at the mystery further, they need far more detailed data, which has led to VORTEX2.

Along with Roger and David, the project’s 30 or so principal investigators include George Bryan and Morris Weisman (ESSL/MMM). David will share field coordination duties with Erik Rasmussen (Rasmussen Systems).

Radar rodeo

David and Erik will have no small task keeping the VORTEX2 armada organized. The experiment will put 40 vehicles on the road, including 10 mobile Doppler radars, from May 10 to June 13. Depending on weather, the team could be anywhere from west Texas to southern Minnesota on a given day. Unlike many field projects, there will be no home base. “We’re planning to leave home on day one and not return till the end of the project,” says David.

The core of the VORTEX armada is its fleet of mobile radars. Several of those will be Doppler on Wheels (DOWs), the NSF national facilities developed by EOL visiting scientist Josh Wurman and his Boulder-based Center for Severe Weather Research. Other mobile radars will come from a variety of collaborators, including the universities of Massachusetts and Oklahoma, the Naval Postgraduate School, NOAA’s National Severe Storms Laboratory, and Texas Tech University.


Vortex study area (left) and schematic of a supercell thunderstorm surrounded by VORTEX2 observing teams (right)

The VORTEX2 study area, shown as a red loop at left, stretches some 900 miles from north to south across the Great Plains. Key sites include Boulder; the National Weather Center in Norman, Oklahoma, which includes the National Severe Storms Laboratory (NSSL); and Texas Tech University in Lubbock. At right is a schematic of a supercell thunderstorm surrounded by VORTEX2 observing teams. The heaviest precipitation, in green, loops around the target area where a tornado might occur.
(Image courtesy VORTEX2.)


The first DOW, built by Josh and colleagues at the Foothills Lab, made its debut as an experimental tool deployed near the end of the original VORTEX. That radar and its younger siblings, now famous through the Discovery Channel series Storm Chasers, have since documented more than 140 tornadoes. With various improvements made in recent years to the DOWs and other mobile radars, VORTEX2 is likely to produce the most detailed radar data ever gathered on tornadoes.

Researchers work on MGAUS, which will be used to launch weather balloons during VORTEX2

EOL’s Bill Brown, Jennifer Standridge, and Tim Lim work on MGAUS (Mobile GPS Advanced Upper-Air Sounding System), which will be used to launch weather balloons during VORTEX2.

Supplementing the radars will be video photogrammetry, which uses frame-by-frame analysis of clouds or debris to deduce the winds swirling around a tornado. In a change of pace from his EOL director’s duties, Roger will take part in photogrammetric analysis; he’ll also participate in aerial and ground-based surveys of tornado damage. “I am a scientist at heart—I can’t wait to get back in the field,” he says.

There’s no airborne radar in VORTEX2, but the project may benefit from an unpiloted aircraft and associated equipment, which together are dubbed an unmanned aircraft system (UAS). Pending flight approvals, the UAS would be deployed by the universities of Colorado, Nebraska, and Oklahoma across western parts of the study area. The UAS would sample weather conditions along critical boundary zones where cool air flowing outward from supercells meets warm, moist air just ahead.

Surface data will come from 10 weather stations mounted on vehicle roofs; StickNets, an array of 24 compact, tripod-mounted stations developed by Texas Tech’s Chris Weiss; and 14 wind-sampling Tornado Pods fielded by Josh. Another key component will be mobile systems for launching radiosondes, including two vans from NOAA’s National Severe Storms Laboratory and two from EOL.

The ultimate moving target

NCAR’s mobile sounding systems have been involved in thunderstorm research for more than 20 years. The current incarnation, dubbed MGAUS (Mobile GPS Advanced Upper-Air Sounding System), can launch up to 15 weather balloons from each van in a single outing.

“The sounding equipment hasn’t changed in the last few years, but our communications needs continue to escalate,” says EOL’s Tim Lim. “Our trucks are full of cell phones, radios, cellular modems, satellite uplink dishes, and moving map displays.”

Tim notes the challenge of orchestrating dozens of vehicles, each charged with a specific task, around a constantly changing and potentially violent thunderstorm. “We have a playbook that looks like it belongs on an NFL sideline,” he says. “Our positions around the storms of interest are hard to judge, and we rely on the field coordinators to help guide us into position and help watch out for our safety.”

There are also staff logistics to consider. Toward the end of each day of operations, VORTEX2 organizers will have to find a town that has at least 50 hotel rooms available for weary observers. As Tim puts it, “It’s hard to get your laundry done when you don’t know what state you’ll be in by nightfall.”

Mobile radars and vehicle-mounted weather stations

The Center for Severe Weather Research deployed mobile radars and vehicle-mounted weather stations near La Crosse, Kansas, on April 1, 2008. VORTEX2 will feature a similar mix of mobile platforms. (Photo by Leora Frankel, CSWR.)

On the plus side, having so many observing tools on the road increases the odds that VORTEX2 will capture the tornado formation cases the team is seeking. To keep things workable, the researchers plan to focus on a single location for at least two hours at a time, putting vehicles in place ahead of a storm that they believe might produce a tornado.

“We’re going to have to roll the dice a bit,” says David. “The real prize will be to document what’s happening in that thirty-minute period before the tornado forms.” The decision-making process will itself serve as a useful test of storm-scale computer models. Such models can now identify a given day’s likely storm type (such as squall lines versus super cells), but they arena’t yet skilled at predicting where or when a specific tornado-producing storm will form. For VORTEX2, NCAR will be supplying special runs of the Weather Research and Forecasting model at a resolution of 3 kilometers (about 1.9 miles).

Daily reports and updates on VORTEX2 will be posted at EOL’s online field catalog, to be coordinated by Greg Stossmeister and linked from the lab’s main site for VORTEX2. Steve Williams will oversee the vast EOL-based archive of forecast and field data. If all goes well, those data will help scientists narrow the lineup of possible culprits in tornadogenesis. The most likely suspects include strong wind shear around the storm in the lowest kilometer and a vertical distribution of unstable air that focuses rising motion in the regions of strongest shear.

With an improved understanding from VORTEX2 of what makes a severe storm produce a tornado, researchers will have a better shot at crafting models that could someday help meteorologists predict a tornadic supercell—and warn residents—before the first hailstone falls.