The residents of Joplin, Missouri, had 24 minutes of warning before a tornado hit their city on Sunday. That gave many people time to take cover, but despite the warning, more than 100 people were killed. Meteorologists would like to be able to warn people earlier and perhaps save more lives, and they have gotten better at predicting the conditions that produce tornadoes. But the swift and chaotic nature of tornado formation might defy our technological capacity to forecast them with greater precision anytime soon.
The National Weather Service and its Storm Prediction Center issue two forms of tornado alerts: watches and warnings. Watches alert people to the presence of storm conditions that breed tornadoes, and can now be called up to five days in advance (10 years ago, warnings came three days in advance, at best). But watches only tell us that tornadoes might be coming—they’re not precise about when or where.
Tornado warnings, on the other hand, are issued when an eyewitness has reported a tornado, or when radar readouts indicate that a tornado is forming. Twenty years ago, residents would get a tornado warning five minutes before the event, on average. Now that time has grown to 15 minutes, on average. But it could be hard to keep extending that range.
Meteorologists already are devoting extensive resources to tornado predictions. They are honing mathematical models of thunderstorms and trying to better understand the conditions that cause those storms to produce tornadoes. The models require abundant data, such as patterns of temperatures and wind flow in the atmosphere, says Roger Edwards, a meteorologist with the National Weather Service. The data come from radar, lidar (light detection and ranging), and remote sensors. Additionally, the National Oceanic and Atmospheric Administration gathers atmospheric data from satellites. Those include geostationary satellites, which hover around 22,300 miles above Earth’s equator and snap images of weather in the Western Hemisphere, and polar orbiting satellites, which monitor changes in temperature, pressure, and other factors.
Back on Earth, meteorologists use many tools to assemble information about atmospheric conditions in tornado-prone regions. Radar and lidar help researchers measure the velocity of winds and thunderstorms. They can employ weather balloons and other unmanned aircraft equipped with sensors that collect data on air pressure, temperature, wind, and humidity.
Some scientists take to the field. In one notable project, the Verification of the Origins of Rotation in Tornadoes Experiment (or Vortex), meteorologists drove around with portable radars and used unmanned aircraft to measure localized characteristics of tornadoes. Their goal was to discover why some vortices of air, or mesocyclones, become tornadoes, while others simply dissipate.
These endeavors have helped researchers develop dozens of algorithmic prediction models for what’s known as ensemble forecasting. In this process, the researchers plug aggregated information into several different models. They analyze the results for hot spots, or areas that the models agree are the most likely for severe weather activity.
As a result, tornado watches are more accurate, and it’s easier for meteorologists to predict where tornado touchdowns are most likely. But only by a little. Meteorologists still can’t effectively measure or model the conditions that immediately precede a tornado. Nor are they sure those conditions exist very long before a tornado forms. Scientists compensate by not issuing many warnings.
“You issue too many warnings, people become complacent. We don’t want to go too far in the wrong direction,” says University of Oklahoma statistical meteorologist Mike Richman.
Richman believes that in the future, warnings could, in some cases, come as early as 40 minutes before a tornado hits. But it will always be hard to get more precise about where the tornado will strike, he and other scientists say.
“These are extreme and small-scale events in the atmosphere that are driven by nonlinear dynamic processes,” says Greg Carbin, warning coordination meteorologist at the National Weather Service. “They are inherently unpredictable.”
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
Data analytics reveal real business value
Sophisticated analytics tools mine insights from data, optimizing operational processes across the enterprise.
Driving companywide efficiencies with AI
Advanced AI and ML capabilities revolutionize how administrative and operations tasks are done.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.