The first time Jana Houser chased a tornado was in 2004, when she was a college student finishing her junior year at Penn State. Houser went out with a team hoping to catch the moment the tornado appeared with a mobile radar. She has been using the tech ever since to better understand how tornadoes form.
Radars send out microwave signals and detect when those waves reflect off objects, and how fast they are moving. Aim a radar at a storm and you can track the distribution and motion of rain and hail, and even insects and debris. Researchers can get a representation of the wind field and how strong the winds of that tornado are. When they pair multiple radars together, they can visualize the inner workings of a storm in three dimensions.
The experience of watching a tornado form right in front of her eyes was exhilarating. Seeing her first tornado was at once the start of a new chapter in her life, and the culmination of another. She’d been fascinated by tornadoes — really, weather in general — since childhood in eastern Pennsylvania. “I would come home from school and watch the Weather Channel like kids watch cartoons,” she says. “I found it absolutely captivating.”
In the early 1990s, she saw footage of a series of major tornadoes during a severe weather outbreak in Kansas and Oklahoma. “I became obsessed with tornadoes and studying them from that point forward. And I really haven't deviated from that path since I was a kid.”
That path has paid off. She became a professor at Ohio State University who uses radar to analyze tornadoes and supercell thunderstorms, including the widest tornado ever recorded. Her research has helped overturn longheld assumptions about tornadoes forming from the sky down. And in 2023, she helped hunt tornadoes for the movie “Twisters.”
Recently, Houser spoke with Sequencer about the (many) open questions that remain about how tornadoes swirl into existence, the challenges of trying to predict the unpredictable, and why she finds tornadoes so irresistible.
The following is a transcript of the conversation, lightly edited for clarity.
What is it about tornadoes that captivates you?
It's this simultaneous fascination and beauty. You can see a tornado in a field, and it almost looks like it's dancing. Almost artistic. But then you see the aftermath of that. There are houses that are destroyed and things that are leveled, and people's lives have been destroyed. Nature is producing this amazing thing that has this violent side as well. It's an unusual paradox.
How would you describe the variety of tornadoes to someone who’s never witnessed one?
Tornadoes take a whole variety of shapes and sizes. You have really thin, delicate funnel clouds. Sometimes they're stationary, sometimes they're moving. They can start out vertically upright and then get stretched out and tilted. You can even have an entirely horizontal component of the funnel. Other times they are huge and wide or they have multiple small, turbulent vortices circling around a central axis.
You can almost feel it in the air...oh my gosh, this is going to be bad.
The colors can be amazing. Oftentimes, you see tornadoes happening close to sunset and you get this really interesting orange hue.
Do you have a favorite tornado you’ve ever seen?
I tend to find beauty in all of them in some way. I've seen a tornado and a rainbow happening simultaneously. That was pretty amazing. That was in South Central Kansas in 2004, in Harper County.
The most destructive tornado I've seen was the EF5 tornado that happened in El Reno, Oklahoma, on May 24, 2011. I watched that one form right in front of my eyes, just spinning and throwing stuff at the ground. Then I saw the May 31, 2013 El Reno tornado which is the widest tornado on record, at 2.6 miles.
In these scenarios, where you get these really big violent tornadoes, there's this sense, and it feels almost physically tangible, like dread, that this is going to be a big one. You can almost feel it in the air. Both the 2011 and especially the 2013 El Reno tornado were preceded for me by just that sense of, oh my gosh, this is going to be bad. Both of those tornadoes became massive very, very quickly. I was operating a mobile radar for both of those. It was almost like an out of body experience.
How often do you go out storm chasing now?
When I was a student, I was out much more. As a professor now living in Ohio, it's much less accessible to me, but I usually find myself out there at least once a year. Every other year, I take students out for an experiential learning course to observe severe convection. This year I was out twice with field campaigns, both for research purposes collecting data.
What have you and others learned from this type of field work?
We can pretty confidently say that most, if not all, tornadoes form from the ground up. The strongest rotation needs to occur at the ground first. But you also need moderately strong rotation in the cloud, because that makes the storm act like a vacuum cleaner, creating an upward acceleration from the surface. This upward acceleration focuses the surface rotation, intensifies it, and pulls it upward, forming the tornado. This process has been confirmed both by high-resolution radar data and high-resolution computer simulations of tornadoes forming.
But the process occurs very quickly. There was a time where we thought the rotation started in a cloud like a funnel cloud does and that the funnel cloud was a visual manifestation of what was happening as the tornado forms. That process sometimes takes several minutes. But as we got better instrumentation, we realized that we're missing a lot with our eyes. Our eyes see the funnel cloud, but what our eyes don't see is the rotation at the ground that hasn't condensed into a funnel cloud yet. This piece was missed in the overall picture of how tornadoes formed until about 15 years ago.
It must have felt great to make such a big leap in the science.
It’s really exciting to change a paradigm and to contribute to that knowledge. I will add that I'm not going to take credit for this idea entirely by myself. There have been a series of studies, both observational and computational that have come together in agreement, over several decades.
It's exciting to do something novel, but what the observations have shown us is also kind of disheartening. From a prediction standpoint, even if we had a hypothetical observation system that was perfect, that saw everything, there's no way we could improve tornado warning time issuance — prior to the tornado forming — because the tornadoes go from nothing to full on tornado fast.
Why is tornado prediction still so difficult, even with better models and radar?
We predict tornadoes quite well in the sense that most tornadoes that happen are warned. The problem is that we overpredict. We over-warn. We have a really high false alarm rate, something like something like 70%.
We also don't really know the cues: When we have collected the rare comprehensive data sets that capture tornado formation, comparing two tornadoes yields different results. Two storms both produce tornadoes, but they have different environments. It's not consistent, necessarily, from case to case.
So how is your research continuing to unpack these fundamentals?
What's driving me right now is trying to answer the question of where that ground-based rotation comes from. I can tell you: you need strong rotation at the ground, and you need cloud-based rotation in a similar location at the same time.
We understand how cloud-based rotation gets there for the most part, but there are a variety of different potential answers to where the ground-based rotation comes from, and it’s not necessarily not the same from storm to storm.
Our eyes see the funnel cloud, but what our eyes don't see is the rotation at the ground that hasn't condensed into a funnel cloud yet. This piece was missed in the overall picture of how tornadoes formed until about 15 years ago.
It seems daunting to go up against so much complexity and uncertainty, especially at scales where we’re talking about mile-wide storms but also microscopic ice crystals that influence them. How do you cope with all of that?
There are a lot of ‘maybes’ and ‘mights’ in my writing. You come up with an answer for one case, and it may or it may not be replicable. That is the nature of a chaotic system like the atmosphere. But as we build on the number of cases, you can gain confidence in a probabilistic sense.
Obviously we have to have some confidence, right? If we didn't then there's no point in doing anything that we're doing. But we have to be careful to not paint with a broader brush than our science and our results really allow.
You can also use observations to tailor additional scientific research. For example, maybe it seems like the air in this one spot of a storm that's producing the tornado is warmer than in the same spot in storms that don't produce tornadoes — so let's load up a whole bunch of vehicles with temperature sensors and send them into this particular area and see if we're right.
Our community consists of individuals with highly specialized expertise. I am focused on radar observations of tornadoes. Somebody else focused on the microphysical characteristics of snow and rain and hail and ice distributions within a storm. Somebody else does computer simulations. We all have a really solid understanding that this is such a complex system and we are very open to seeking outside, physically justifiable perspectives to help us answer our own questions.
