“How’s the weather?” may be small talk for most people, but for Howard Bluestein the weather has been a subject of fascination since early childhood. “I was interested in the weather wherever I was,” he says.
Bluestein took up meteorology and in 1976 joined the faculty at the University of Oklahoma, where he focuses on severe storms. (See also the story in Physics Today, August 2018, page 22.) His NSF-funded research involves hopping into a truck to chase tornadoes. Even after roughly four decades of tornado research using increasingly sophisticated radar systems and other tricks, Bluestein says the burning question driving his study of twisters remains: “Can you predict which storms will produce tornadoes and which will not?”
PT: How did you get interested in severe weather?
BLUESTEIN: When I was five years old or so, in 1953, there was a big tornado in Worcester, Massachusetts. I remember playing outside in the street, and my mother told me to come inside, that tornadoes suck little boys up into the air. I didn’t believe it, but I did go in the house.
One day, lightning struck the TV antenna on the roof of our house and the TV blew up. That got my attention. We also had a lot of good snowstorms, and they always excited me.
Then we moved to South Miami Beach, Florida, where I lived from the second to the sixth grade. The weather was totally different from what it was in the Boston area. Thunderstorms would come up along the sea-breeze convergence line, which was just inland, but the storms would never move eastward over us. That was kind of tantalizing.
PT: Was meteorology an obvious career choice for you?
BLUESTEIN: I didn’t know that you could do research in meteorology. At MIT I decided to major in electrical engineering because it combines physics and math.
When I was a senior I hadn’t decided whether to go on to graduate school in double-E or to think about meteorology. I realized I was spending a lot of time on the telephone calling up the weather. There was no internet. I had to indulge in this obsession using a telephone. Then I spent a summer working for Fred Sanders, who was a professor of meteorology at MIT. I realized, “My god, you can have a career and you can do research and you can enjoy the weather.”
PT: What did you study?
BLUESTEIN: MIT made an offer I couldn’t refuse: They gave me an assistantship in meteorology and told me I could stay in double-E. My master’s involved predicting cloud patterns using the relatively new satellite imagery. I made connections with the National Hurricane Research Laboratory in Miami—now it’s the hurricane research division of NOAA. I went there for several summers, and it led to my PhD in tropical meteorology.
PT: How did you end up at the University of Oklahoma?
BLUESTEIN: The director and founder of the National Severe Storms Laboratory in Norman, Oklahoma, was at MIT on sabbatical while I was finishing my thesis. He saw that I was interested in severe weather and said, “Why don’t you come study severe thunderstorms and tornadoes?” I said, “Gee, that sounds interesting, but I wouldn’t go to Norman, Oklahoma, if it were the last place in the world.”
I was flown down to OU for an interview. And I saw that there was civilization, lights, running water, and it was a nice place. It wasn’t the dust bowl I thought it was. I took the job. I was just going to stay for a few years and then move on, but it’s been more than 40 years.
PT: Describe your early work in Norman.
BLUESTEIN: When I moved here, in the mid 1970s, the National Severe Storms Laboratory was just starting to use Doppler radar to study the weather. At the same time, people were starting to chase storms. I decided it would be fun to do that. We were providing “ground truth” for the people back at the National Severe Storms Laboratory who were looking at the storms remotely.
We would go out, and they would say, “We see a mesocyclone signature on the radar. Do you see a tornado?” Or “We see a signature for hail. Do you see hail?”
This was very satisfying and a lot of fun. After chasing storms for a year or two, I realized we could safely get near tornadoes and severe thunderstorms. So I had the idea of placing instruments in the path of storms. In 1980 I was introduced to Al Bedard at the Wave Propagation Laboratory in Colorado. He said he could design an instrument that I could take out and try to put in the path of tornadoes. We called the instrument TOTO—for Totable Tornado Observatory—in deference to Dorothy’s dog in The Wizard of Oz.
This instrument could measure temperature, pressure, wind direction, wind speed, and humidity. It was a 400-pound instrument carried on the back of a pickup truck. We would put it in front of a tornado and the tornado would disappear, or it would change direction. We did this for several years. It was just too difficult.
PT: That sounds a lot like the machine used by the fictional meteorologists in the movie Twister.
BLUESTEIN: Several of the people who made Twister visited me in Norman. I showed them work we had done using TOTO and gave them some journal articles. Lo and behold, they took the device we were using, renamed it Dorothy. They asked if I’d be a consultant on the film, but they wanted me to be completely committed. I didn’t have time.
PT: What other technologies have you used?
BLUESTEIN: In the late 1980s, my students and I started using a frequency-modulated continuous-wave Doppler radar. Instead of sending out pulses, it sent out a continuous beam.
What we did was mix the frequency of the radar with the frequency of the signal coming back. We listened to the beat frequency signal through stereo earphones. In one earphone you could hear the receding pitches—the frequency would get lower. And in the other earphone you could hear the approaching pitches, for which the frequency would be going up. The higher the pitch, the higher the wind speed coming toward you.
We were able to go out with this radar, get really near tornadoes, and for the first time ever get several miles from a tornado and actually look at the winds near the ground, within 100 meters.
Around 1990 one of my colleagues, Josh Wurman, decided to put a pulsed Doppler radar on a truck. He decided to go to X-band, which is 3 centimeters wavelength. We decided to use a W-band, or 3 millimeters wavelength, radar. The reason we went to 3 millimeters is that the antenna has to be smaller than the truck, and you don’t want it to hit trees. And when you go get gas, you don’t want the antenna to blast through the roof of the gas station.
It turned out that we could get very high spatial resolution. At X-band, the antenna resolution is typically 1 degree, whereas at W-band, we got to less than two-tenths of a degree. We went out with this radar for a number of years and made fine-scale measurements of wind speeds in tornadoes. Josh did the same with the X-band radar, but he was seeing the whole storm.
PT: Are you still using mobile W-band for chases?
BLUESTEIN: Eventually we began to work with a polarimetric radar. The radar sends out two beams, one that is horizontally polarized and one that is vertically polarized. If you have rain falling, the large raindrops tend to flatten as they fall, so they become wider than they are tall. You can tell by looking at the ratio of energy you get back from the two beams whether you are looking at large raindrops or small raindrops—the small ones aren’t flattened.
You can do other tricks too. When you look at a tornado, you can actually tell if you are looking at debris flying through the air or looking at rain or hail. This is very useful. You know from looking at the winds whether there’s a vortex there, but is the vortex actually at the ground kicking up stuff? With a polarimetric radar you can tell that.
Around 2007 or so, we realized that in supercells—long-lived thunderstorms with a rotating updraft—tornadoes form really quickly, in a matter of seconds. Most radars that scan storms take a few minutes to scan the complete storm, so they can easily miss the formation of a tornado. We hooked up with someone who had a Navy rapid-scan radar, and we were able to actually follow tornadoes forming and decaying inside storms. After a few years we decided we wanted to have a polarimetric radar that could scan rapidly.
A colleague did this with frequency hopping. You take the radar, and instead of transmitting at one frequency, you change the frequency just enough that it’s like having a completely independent radar. We have taken this radar out in some major tornadoes.
PT: How do you prepare for a storm chase?
BLUESTEIN: Nowadays, I look the day before at computer forecasts of wind speed, wind direction, moisture, and so on. That gives me an idea of whether the conditions are going to be suitable for supercells and possibly tornadoes. When I get up in the morning I look at the latest real data and computer models. I look at runs from ensemble models, models run 10, 20, 50 times with slightly different initial conditions—you never know the actual conditions exactly. Based on the results, you pick your target area, and you try to drive out there before the storm has formed.
We try to place ourselves in a position just ahead of but a little out of the way of the storm. I don’t like to get closer than 3 kilometers. And then we start scanning the storm with the radar.
PT: What have been some of your most important discoveries?
BLUESTEIN: I think we were the first to measure wind speeds in an F5 tornado, with a Doppler radar. This was back in 1991. The wind speed was 125 meters per second. Then in May 2013, in the El Reno tornado in Oklahoma, we tracked 24 different sub-tornado-scale vortices rotating around the tornado. They have extremely high wind speeds, some as high as 300 miles per hour [134 m/s] relative to the ground. And we have documented many anticyclonic tornadoes and found that they tend to come paired with cyclonic tornadoes or mesocyclones.
PT: Why is measuring the wind speed important?
BLUESTEIN: People estimate—they still do—the wind speed from the damage, not from actual wind measurements. So they might estimate what forces are needed to lift a roof off a house or a house off the ground.
PT: What are you working on now?
BLUESTEIN: We are still trying to discover the structure of tornadoes near the ground. It’s especially difficult right above the ground—the radar cannot get to the lowest 10 meters, where the damage gets done—and it turns out there are some interesting dynamical effects near the ground because of friction. I’d like to use mobile Doppler LIDAR [light detection and ranging] along with mobile Doppler radar, so we can have the radar scan the whole storm and the LIDAR focus on the tornado.
And I am trying to arrange a collaboration with NASA. They would fly an aircraft with a radar mounted to look downward over a storm, while our ground-based radars would look up at the same storm. We also want to use the latest satellite radar data in conjunction with our ground-based data and NASA’s airborne radar. Are there clues from looking at the top of the storm that will tell us if there is going to be a tornado near the ground?