Researchers have combined data collected by scientists and an IMAX filmmaker for the first detailed look at  what happens in the bottom 100 feet in a tornado.

Tornado Intercept Vehicle Tornado Intercept Vehicle, designed by filmmaker Sean Casey.  Inside is a camera used to create an IMAX movie of a June 2009 twister in Wyoming.  (Tornado Alley, LLC)

In an article to be published in the June issue of the Bulletin of the American Meteorological Society (BAMS), Joshua Wurman, Karen Kosiba and Paul Robinson of the Center for Severe Weather Research in Boulder, Colo., say: “This has been a holy grail for us for years since it  has been so difficult and rare for us to get simultaneous high resolution (radar) data and surface data in a strong tornado.”

Since the 1990s scientists have been using Doppler weather radars on trucks to capture close-up views of winds in and around tornadoes. But these radar observations have mostly missed the bottom 100 feet or so of tornadoes because researchers like to keep the mobile radars a mile or so from tornadoes where hills, trees and buildings interfere with low-level radar waves.

While scientists involved in the VORTEX 2 tornado research project were busy collecting Doppler radar and other data on the tornado that hit Goshen County, Wyo., on June 5, 2009, Sean C. Casey was driving  his Tornado Intercept Vehicle (TIV) through the twister, filming the climax of his “Tornado Alley” IMAX movie.

Related: Tornado data could help forecasters increase warning times and reduce false alarms

In their examination of the bottom of the tornado, the researchers used VORTEX 2 data plus the TIV’s IMAX movie, other video, still photos, and wind and other meteorological data from the TIV’s instruments.

“Combined TIV and [mobile doppler radar] observations provided the unique opportunity to integrate radar, in situ wind and video data to address outstanding questions related to the low-level tornadic winds,” the researchers write.

The TIV video is especially valuable because it shows exactly when and how the tornado broke some of the electrical-wire poles along the road the TIV was using.

“Up until now, there has existed no direct inter-comparison between anemometer-measured winds and real-time documentation of damage within a tornado,” Wurman says.

The outer, gray “shell” is the tornado’s debris cloud. Red is the area of fastest winds. Yellow is an area winds slowing to nearly calm at the center. Black arrow is the tornado’s direction of movement. The curved arrows represent typical winds in 3 dimensions. The line the tornado is crossing is the north-south road the TIV used. The outer gray “shell” is the tornado’s debris cloud. Red is the area of fastest winds. Yellow is an area of winds slowing to nearly calm at the center. Black arrow is the tornado’s direction of movement. The curved arrows represent typical winds in three dimensions. The line the tornado is crossing is the north-south road traveled by the TIV. (Center for Severe Weather Research)

In brief, the researchers found that the tornado had spiraling inward and upward flows near the ground that were met by a downdraft coming down from roughly 3,000 feet high to below 300 feet but not all the way to the ground.

The Center for Severe Weather Research scientists are now analyzing data from a tornado that hit their Rapid-Scan Doppler on Wheels (DOW) and a mobile-mesonet (a pickup truck with attached meteorological instruments) near Russell, Kan., on May 25, 2012 with no injuries to the three people in the DOW nor the three in the mobile mesonet, including Wurman.

Scientific storm chasers strive not to get dangerously near tornadoes, but in Kansas, “we didn’t drive away because the tornado formed very near the road, and was on us before we could react. We didn’t expect it to form when and where it did,” Wurman says. “The storm structure was less “classic” than the more predictable one in Goshen. “If  we had known that this was going to happen we would have not gotten a good data set; we would have backed off.”

The tornado destroyed a house approximately 100 feet from the DOW and its debris broke one of the truck’s windows.

As the tornado headed toward the truck its radar was collecting good wind speed and direction data from roughly 11 feet above the ground up to around 240 feet, “as the crew cowered inside,” Wurman says. In addition to this DOW,  two other Doppler radars, one about a half-mile away, also recorded the tornado’s winds.

This data will help fill in the picture from the Goshen tornado, Wurman says.

“We should be able to really resolve the shallow inflow layer, how much and how fast air is going in or out of the tornado and how this relates to tornado evolution. We will be able to quantify how high above the ground the strongest tornado winds are,” Wurman says.

Doppler radar and other data from the Goshen and Russell tornadoes as well as similar data in the future would help scientists create better computer models of tornadoes.

In addition to improving the understanding of tornadoes and how they destroy structures, such models could help design buildings that tornadoes wouldn’t damage as severely.

Such models could help improve estimates of tornado wind speeds. Since tornado wind speeds are hardly ever directly measured, the National Weather Service examines tornado damage afterwards and uses the Enhanced Fujita (EF) scale to estimate speeds based on the severity of the damage.

Wurman notes that the scale is based on several assumptions about the relation between wind speeds and damage. Both the Goshen and Russell, Kan., tornadoes illustrate limitations of the EF scale.

Based on damage to the electrical poles (the only “structure” hit) before any Doppler data were available, the National Weather Service rated the Goshen tornado as an EF-2 with speeds of 111-135 mph, but the DOW measurements showed its speeds were close to the 165-166 mph EF-3/EF-4 boundary.

A separate NWS office also rated the Russell, Kan., tornado as an EF-2 based mainly on its destruction of the house that sent debris into the DOW truck.  Here, the highest speed recorded by the mobile mesonet when the tornado hit it was 96 mph, which  would make it an EF-1 twister.