This article is from the In-Depth Report The Science of Tornadoes

Droning It In: Storm-Chasing Unmanned Aerial Vehicle Makes First Foray into Nascent Twister

Unmanned aerial vehicles hold great potential in their ability to provide crucial data about tornado behavior--if only the FAA allowed scientists to use them

University of Colorado at boulder

Unmanned aerial vehicles (UAVs) are no longer just gizmos in a geek's garage or military tools that fly reconnaissance missions considered too dangerous for humans. They are increasingly being used for scientific study. And this spring, a UAV dedicated to research science made aerial history.

On May 6, a diminutive aircraft called the Tempest was the first official UAV to intercept a supercell thunderstorm, the type of storm that produces tornadoes. The aircraft and its crew of engineers from the University of Colorado at Boulder (C.U.–Boulder) and meteorologists from the University of Nebraska–Lincoln are a critical part of an armada of 100 storm-chasing scientists conducting the largest study of tornadoes in history. The two-year field experiment known as VORTEX 2, running from May 1 through to June 15 this year, will help scientists better understand when and how tornadoes form. Teams travel across the Midwest in tight formation chasing and surrounding tornadic storms to measure wind speed, temperature, humidity and pressure using mobile radar trucks, anemometers, disdrometers and balloon launchers.

Added to this list, the Tempest is designed to take center stage, flying into the rear flank of supercells, 150 to 300 meters  aboveground—a sweet spot for gathering data that are inaccessible to the current instruments or any manned aircraft. Now that it has made its maiden data-gathering flight, the Tempest stands ready to continue this work for the duration of this tornado season.


But it's not that simple: Federal Aviation Administration (FAA) regulations restrict the Tempest from freely flying throughout the range of "tornado alley"—the area in the U.S. between the Rocky and Appalachian mountains. Currently, the researchers can only work within about 58 small grids of land, each about 1,000 square kilometers in area in northeastern Colorado and portions of Kansas and Nebraska. Getting flight clearance in just these areas required three years of coordination with the FAA. The researchers had to file 60 applications in order to be waived from having to comply with all the rules specific to manned aircraft.

When Brian Argrow, director of the University of Colorado Research and Engineering Center for Unmanned Vehicles, announced that he wanted to apply for the entire area stretching from central Nebraska to the Texas panhandle—an area that sees a concentration of twisters—FAA members responded with an audible gasp. Currently the FAA has a freeze on further applications to fly.

The agency's resources are simply overwhelmed and UAVs do not fit neatly into existing procedures and protocol.

"The issue is that as of today there are no formal rules set for unmanned aircraft, only rules for manned aircraft," Argrow says. The technology has leapt ahead of bureaucracy and the FAA finds itself asking UAV flyers to comply with rules that border on comical in this context, such as ones dealing with seat belts.

Of course, the FAA requirements are for security and safety: Terrorists or anyone for that matter could hijack a UAV, and a UAV increases the chances of midair or even ground collisions with other aircraft, buildings or humans.

Still, the battery-powered Tempest is relatively tiny, about three meters long with a 3.2-meter wingspan.* It can fly for an hour on one charge at up to 130 kilometers per hour. Researchers hand launch and guide it into areas three to five kilometers away from a developing supercell where it can handle up to 95 kph winds. Sensors on the Tempest's body measure pressure, temperature and moisture in a storm's downdraft inflow—the two airstreams that feed a tornado—and the UAV transmits in-flight data back to the team just in case it gets ripped apart by sudden gusts.

What is the risk?
It is tough at this point to pin down the real risk that UAVs present, but they certainly do not maneuver like conventional aircraft.

"Currently, no technology allows these planes to sense other aircraft and make turns to avoid it," says Les Dorr, an FAA spokesperson. The single most important criterion for any unmanned plane is to be able to "see and avoid" other aircraft, he adds.

UAVs may only weigh 5.5 kilograms but can still cause serious damage if they collide with piloted aircraft. For instance, the Canada geese that flew into the engines of US Airways flight 1549 resulting in an emergency landing on the Hudson River in 2009, weighed about 8 kilograms each.

Still, some VORTEX 2 scientists wonder just how high the risk is for a UAV to collide with another aircraft. "There will never be a sane human pilot in that area of the storm," says Erik Rasmussen the co-lead principal investigator for VORTEX 2. "I suppose a nutso crop duster guy could fly, but the odds are so miniscule relative to how many hoops we must jump through to get permission."

In order to get permission for Tempest flights the FAA required that  during flight continuous visual and radio contact be maintained with the UAV, so the team built a computer that instructed the drone to follow a vehicle on the ground below it. "[It is] a bit ironic that we have an aircraft whose primary limitations are the capabilities of a ground vehicle," Argrow wrote in an e-mail.

The FAA had allowed for a manned "chase" aircraft to accompany the UAV. "But we didn't want people to be that close [to a tornadic storm]. It's dangerous," says Jack Elston, a graduate student in aerospace engineering at C.U.–Boulder. After all, "unmanned" is the entire point.

Last-minute flight maps
The FAA also demands a preplanned flight map submitted 48 to 72 hours prior to launching the Tempest. The irony of this requirement is painful: If the National Weather Service's Storm Prediction Center could provide 72 hours notice of where a tornado will form and where it will travel, drones would not be needed to study tornadoes in the first place. Right now, the best meteorologists can do is provide a 13-minute advance warning. This is the very reason VORTEX 2 and the Tempest UAV exist—the project's goal is to extend the warning times.
But when Argrow and his team started operations in early May, the FAA air traffic controllers quickly learned that their mandated rules simply did not work with a UAV tasked to catch fast-developing supercells. The Tempest team had been forced to submit numerous flight plans simultaneously—one for each of the 10 or 20 permitted land blocks they might cover while tracking storms 72 hours out—and this overwhelmed the controllers. The FAA realized they would need to reduce the 72-hour lead time, so the flight plans could be based on more recent meteorological information and thus be more precise, asking for access to only two blocks of land at any one time.

Still the scientists, like spiders waiting for insects to fly into their web, have had to wait for storms to form within the FAA-cleared zones and hope that a storm remained within those blocks. If storms keep skirting the allowed sectors, the scientists may become increasingly frustrated with their progress this season, Argrow says. If the Tempest were free to coverthe entire VORTEX 2 domain covering more than 10 states, they would be able to intercept all the storms other the teams have targeted, which could have potentially jumped their intercept frequency from two storms in one month to maybe two in one day.

Uncertain future for UAV-collected data
Researchers within the VORTEX 2 team say they have hope for UAVs but wonder if they will ever get past FAA regulations so that they can roam freely through the Great Plains.
"It's understandable that some are skeptical. We've been struggling since 1996 to get a UAV into a storm," Rasmussen says.

Frustration increased last summer when a group connected to Discovery Channel’s Storm Chasers flew an unmanned plane straight into a forming tornado. When the scientists learned this, they were despondent. "We've been following the rules, and here's someone breaking the rules and could potentially ruin it for us," Rasmussen says. If something had gone wrong with that plane, it's certain the FAA would have further delayed anyone flying anything into a storm, he said.

Meanwhile, back in April 2009, the FAA set up a committee to develop formal recommendations specific to unmanned small aircraft. This would relieve the burden of submitting countless applications to fly.  But, at best, the FAA does not expect to have the new regulations confirmed until 2012, Dorr says. Regardless, Argrow says the new UAV-specific guidelines will have little to no impact on his team's research. There is a chance the guidelines may only apply to flight tests at heights below 1,000 feet, not the sort of free rangeArgrow and his team need in order to study twisters. For storm-chasing scientists expanded regulations that permit clearance to fly UAVs can't come fast enough. Although ground-based mobile radar can measure wind speed and direction, it cannot measure thermodynamic variables such as temperature, pressure and humidity. Such variables can only be measured by placing the instrument in the locus where the measurement needs to be taken, 300 meters up. The Tempest may be the only chance of measuring these variables in a developing storm.

"These [UAV] planes will be incredibly valuable. For every 10 million observations of winds, we have only one for temperature, pressure and humidity," Rasmussen says. Not only will they sample air that cannot be studied today, but the UAVs are the one chance we have of balancing out the data, he adds.

For now the Tempest is on call for the next supercell to wind its way into a section of the 58 FAA-cleared grids. Luckily, the forecast is calling for storms to migrate into this domain and the Tempest is ready to meet up once again with the VORTEX 2 armada. "It's looking good. We should be very active for the next week and through to the June 15 end date of the project," Argrow says.

*Correction (6/9/10): This sentence was edited after publication to correct the wingspan length .

Rights & Permissions
Share this Article:


You must sign in or register as a member to submit a comment.
Scientific American Back To School

Back to School Sale!

12 Digital Issues + 4 Years of Archive Access just $19.99

Order Now >


Email this Article