Tornadogenesis
When things get rough down low
Tornadogenesis is a compound word that MS Office, Grammarly, and ChatGPT hate. I imagine it is like nails on a chalkboard to their algorithms. They see it and want to break it up even though it is real. Like paparazzi and celebrity marriages.
Tornadogenesis describes the process of tornado formation. They occasionally form in long, continuous squall lines or in isolated young storms that have not yet begun to rotate. But the most destructive tornados form in supercells – strong thunderstorms with an updraft.1
In a prior newsletter, we discussed the boundary layer — a region of air near the surface. Currently, most computer models simulating tornadogenesis assume a boundary layer that lacks turbulence. That is, the winds tend to consistently move more horizontally than vertically.
In reality, boundary layers in supercells are known to be highly turbulent. So, why do simulations often assume otherwise? Blame computer engineers plus scientific inertia.
It’s really just been lack of computing power, and now that we have it, most people have just kept simulating storms the old-fashioned way, ie with laminar boundary layers. I guess it’s just been assumed that the gory details introduced by including a turbulent environment don’t change the fundamentals. I had always assumed that myself, in fact (until now).” – Dr. Paul M. Markowski, Pennsylvania State University
But things change, and in a 2020 paper in the Journal of the Atmospheric Sciences, Dr. Markowski described one of the first tornadogenesis simulations to incorporate boundary layer turbulence. It starts by adding small, random temperature differentials to the simulation space. This sparks up-and-down movement in the air (turbulence2). The amount and type of turbulence his simulation created were similar to what may be found in the early evening hours when tornadoes more often form. He also used computer models with higher grid spacing than most (75m), achievable thanks to continued increases in computing power.
He ran 25 different simulations with this new system, all of which generated tornadoes. One of the surprises was the variability in the types of tornadoes generated with this new simulation. They were quite different in strength (EF0-EF3), duration, and in other ways. It’s a sign of the critical role of small-scale environmental features in tornadogenesis.
One of these is not like the others
Of those 25 simulations, one (#9) stood out as different. In a 2024 paper, Dr. Markowski reran that particular simulation for a deeper look. The only difference was that he set it up to output more data. One of #9’s unique aspects was the influence of ground-based eddies in the surface wind. They served as seeds for tornadic production as the supercell’s updraft overran this turbulence. This hasn’t been seen before in a simulation.
It could have consequences for tornado forecasts. Given that the eddies are shown to be key to tornado formation in the simulation, forecasters could look for environments in which significant coherent structures like those in the simulation would be more favored. These would be environments with neutral surface-layer stratification and high ground-relative wind speeds. So, windy early evenings, basically.
Simulated data is just that – simulated. Scientists often use them to better design real-world data collection, which, in turn, is used to confirm and adjust the models behind the simulation. It’s a closed-loop system, like politics. In this case, most prior observational studies of tornadogenesis have not been designed to detect the kind of environmental differences this simulation predicts. Markowski suggests ways of collecting more data to test the simulation results, perhaps by using small, balloon-borne sensors in supercells (swarmsonde). This is the only way to compare real atmosphere data to the models.
Dr. Markowski’s research group invented a swarmsonde technique to fly 60-80 probes through a single storm. The top image is one such probe, and the bottom is a collection of data from one swarm release. Check out the overachiever that got to 8km! (Courtesy Dr. Markowski)
If this finding proves to be robust over time, it could reshape our comprehension of tornado formation within their most notorious setting—supercell thunderstorms—by emphasizing the critical importance of environmental turbulence. This research not only deepens our theoretical understanding of tornadogenesis but also has profound implications for improving tornado forecasting and enhancing the collection of observational data.
And now for something completely different…
Writing headlines is hard - especially when what you want to say conflicts with reality. Check out this headline from a Disney tourism site: “10 Hurricanes to Hit Disney in 2024”. Written like a journalism article, it includes this choice quote from an interviewed meteorologist: “This isn’t your standard ‘X’ number of storms style forecast…”. Except, of course, when it is.
We acknowledge additional contributions from Paul Markowski.
A plume of buoyant air that spins as it rises
The difference is that the turbulence is in the environmental conditions of his simulations. In most prior simulation techniques, turbulence existed in the storms, of course, but not in the initial environment.