Scientists are getting better at predicting tornadoes, which helps save lives. But do people always heed the warnings?
A cluster of tornadoes that left eight people dead, destroyed scores of buildings, and left thousands without power in Illinois, Indiana, Kentucky, and nine other Midwest states this past weekend had at least one silver lining: They didn't kill as many people as they could have, according to local media.
Kentucky.com credited forecasters' "uncannily accurate predictions," which were combined with television and radio warnings, text-message alerts, and storm sirens. "[It] almost certainly saved lives as rare late-season tornadoes dropped out of a dark autumn sky," wrote the paper.
The first accurate tornado warning was issued back in 1948, when observant meteorologists at Tinker Air Force Base in Oklahoma noticed that weather conditions seemed similar to skies that had birthed a twister just five days earlier. Since then, scientists have used weather balloons, radar, and other measurements to improve the science of storm prediction.
Predicting tornadoes is still harder than forecasting hurricanes, since tornadoes are much smaller and take more unpredictable courses. But the science is evolving.
National Geographic spoke with Charlie Neese, a severe weather expert and television meteorologist based in Nashville, Tennessee, about tornado prediction science and safety.
A lot of media are saying that the death toll from this weekend's storms in the Midwest could have been much worse if it hadn't been for advance warnings of severe weather. Do you agree?
Oh, yes, absolutely. For the size of the tornadoes that struck Illinois, Indiana, and Kentucky and the number of people affected, there were relatively few deaths, which means to me people were aware of the threat and they took the proper action. The key to survival in a tornado is knowing that you need to seek shelter.
I know forecasters were saying as early as Thursday that there could be severe weather by Sunday. On Friday they increased that messaging, and by Saturday there was wide awareness—awareness that definitely saved lives. Then it's up to people to take that information and then prepare.
What if people don't act on the warnings?
That's still where there is a missing link and a disconnect. For some it does connect, but some people will hear the warnings and they won't do anything about it. They think it won't happen to them.
The more severe weather stays in the media, the more it will remain top of mind, and that helps. Lots of people watched Hurricane Sandy on TV, even if they didn't live on the coast. In May, when tornadoes hit Oklahoma, lots of people watched. People internalize that. It makes them feel like they should prepare, so when they have a risk in their own area they can respond.
Do people get enough notice?
Our prediction system does a very good job of letting people know a few days in advance that severe weather may be coming.
Do you think the National Weather Service's current program of issuing a tornado "watch" when a storm is likely, and a "warning" when a storm is imminent, is working? Some critics have argued that people are likely to tune out the watch part.
I know there is a lot of research going into the effectiveness of the system as it stands right now. After so many people were killed in Alabama on April 27, 2011 [65 people in the Birmingham and Tuscaloosa areas], the National Weather Service started working with social scientists at the University of Alabama-Huntsville to interview people in the affected areas. They asked why (or why not) they took action, and when.
The social scientists found that on April 27, when people heard the watch issued, a set of people took it very seriously and stayed tuned to weather info. So when the warning was issued, they already knew what they were going to do and they took immediate action.
But a much larger portion takes a watch, puts it in the back of their mind, and they don't necessarily change their day-to-day plans. So when a warning is issued, they don't take immediate action. They wait for confirming information. They wait to hear a tone in a meteorologist's voice that expresses extra danger, or they want to see the tornado on TV, coming toward them. Or they want to open their windows to see the weather outside, or get a call from a neighbor who is downwind.
What we're trying to do is create enough time between a warning and an event taking place so people have enough time to get safe. It used to be just four to seven minutes after a warning was issued [before a storm hits], but now it's common to have ten minutes or significantly more [in Moore, Oklahoma, in May, some residents had 16 minutes after a warning was issued before a tornado struck].
Our warning system is getting better, but we have to figure out how you get people to take the warning seriously. Every second that you delay action you increase risk of death and injury.
How could warnings be improved?
One thing the National Weather Service is experimenting with is being more descriptive in describing storms. So instead of just giving wind speeds, saying that trailer homes could be overturned, or that homes could be significantly damaged, for example.
Computers continue to get faster and more powerful, which can help scientists make better weather forecasts. What other technological advancements have improved tornado science?
Doppler radar continues to get better and better, and gives better resolution of wind. Dual polarization radars are being used, which heighten the ability to see rotation and see if there is debris in that circulation. It can identify sizes of particles, so you can identify debris. If you see a lot of debris, it gives you a much better idea that the storm is on the ground.
Computer models used to forecast weather are getting better. The algorithms are getting more advanced. In the past you could forecast a general area where supercell thunderstorms [which spawn tornadoes] could develop, but some of the models we use now have higher resolution. This will lead to one day, I believe, being able to forecast possibilities of a tornado hitting a particular neighborhood.
Computer modeling is where a lot of the energy is going—that and Doppler radar. Scientists take data out of Doppler radar and run it through specific algorithms. One looks at rain, one wind, one debris, et cetera. It helps us better understand the weather.
Was there anything else notable about the recent storms?
While spring tends to produce more tornadoes, they're not necessarily uncommon in fall. It's always good to be prepared.
This year we have had a low number of tornadoes overall so far, but the number of violent tornadoes has been high. The national average is 1,200 tornadoes a year, and we have had a little more than 800, but a high percentage of those have been violent ones.