Column: Five myths about tornadoes
The scene in Moore, Okla., this past week was hauntingly familiar. The images of cleanup crews picking through the wreckage of two elementary schools transported me back to 1957, when an F5 tornado struck my Kansas City neighborhood, destroying my kindergarten and leaving 44 people dead. Thankfully, we’ve learned a lot since then that can help limit tornado casualties. But many misconceptions persist — misconceptions that can encourage bad policy and put lives at risk. I’d like to dispel some of the myths.
Meteorologists aren’t any good at forecasting these storms.
How does 99.3 percent sound? In 2011, 553 people lost their lives in tornadoes. For all but four of those victims (99.3 percent), both a tornado watch and a tornado warning were in effect before the storm arrived.
Modern tornado warnings are Nobel Prize-worthy endeavors that combine weather science, social science and technology. As recently as 1990, people in the path of a tornado were lucky to get five minutes’ warning. Now, thanks to advances in radar, computer simulations and research on how tornadoes develop, the average “lead time” is 12 minutes — and more than 15 minutes for major tornadoes. The city of Moore had a stunning 36 minutes of warning.
In addition to the explicit warning to take cover, the National Weather Service issued a tornado watch more than two hours before the tornado arrived in Moore, allowing people to move their valuables into storm shelters or even drive out of the area. There were also tornado “outlooks” four days before the Moore tornado. Those stated, in words and graphics, that central Oklahoma had an elevated risk of major tornadoes on that Monday.
The one problem that weather science needs to address is false alarms: For every four warnings issued, only about one tornado touches down.
Those false alarms can cause people to question the credibility of the warning system. That said, if a significant tornado is headed for your area, the chance of an advance warning is excellent.
Warning systems don’t work.
Since Weather Bureau civilian tornado warnings (as we would think of them today) began in 1957, there has never been a tornado that claimed more than 100 lives — with one notable exception.
On May 22, 2011, an F5 tornado struck Joplin, Mo., population 50,000. This was one of the rare times when almost everything went wrong with the warnings. The National Weather Service misreported the location and direction of the tornado. The sirens were not sounded in a manner consistent with the warnings, leading to confusion. And the tornado was enshrouded in rain, so people couldn’t see it. One hundred sixty-one people died.
A week ago Monday, a tornado of equal strength and larger physical size struck Moore, population 55,000. It was similarly difficult to recognize along its path because of rain and debris. Yet the warnings went out as they were supposed to, and Moore experienced one-seventh the number of deaths in Joplin.
The tragedy in Joplin was the exception that proves the rule: Tornado warnings are extremely effective at saving lives. By some estimates, they save 1,000 to 2,000 lives a year.
The earlier the warning, the better.
“Anything that gives us even a few additional minutes is extremely valuable,” John Trostel, director of the Severe Storms Research Center at Georgia Tech University, told The Washington Post this past week.
That was true in the days of five-minute warnings. The idea that there could be too much lead time was unfathomable when the safety rules were first conceived. But research has shown diminishing returns for lead times beyond 15 minutes.
It takes only a few moments to dash down to a basement or into an interior closet or bathroom, as the National Weather Service recommends. Too much lead time tempts people to lose focus on safety: Should I leave the basement to get that photo album? Should I go out to put the car in the garage?
While anecdotal, there is also considerable evidence that people will stay sheltered for only so long. In one episode in Kansas in 1991, two preteen girls who were home alone had ordered pizza before the sirens went off. After they’d spent about 15 minutes in the basement, the doorbell rang. Since nothing had happened and the pizza guy was there, they figured it was OK to come out. They went back upstairs and paid for the pizza. As they were doing so, the cat ran outside. They chased it around the side of the house — and saw the tornado just about on top of them. Everyone ended up safe. But they certainly put themselves in unnecessary danger.
Community shelters are a bad idea.
The city of Moore’s website makes the typical case against community shelters. “The three places where you shouldn’t be during severe thunderstorms are in mobile homes, in your car, or outside,” it says. “Going to a community storm shelter puts you in two of these places.” The site adds that there isn’t enough lead time for people to get to a shelter and that housing a population of more than 55,000 “would require building something like an underground sports arena.”
Again, that sort of thinking harkens back to the era of five-minute warnings.
In situations when there’s more than a half-hour of warning, the possibilities are very different. Community shelters are among the safety ideas that need to be revisited.
Sure, a single shelter for 55,000 wouldn’t be practical, but a city could consider multiple public shelters. Andover, Kan., population 12,000, has two, and there were no issues with people reaching the shelters and hunkering down there when a small tornado hit the town last year. Alternatively, employers could be encouraged to open their spaces — which tend to be stronger than homes — for their workers’ families.
Rethinking safety recommendations is especially important in places like Moore, where many houses don’t have basements, and in areas with many mobile homes.
Climate change is producing tornadoes of increasing frequency and intensity.
There have always been F5 tornadoes, and we will continue to experience them regardless of whether the Earth’s temperature rises or falls. National Weather Service figures show, if anything, that violent tornadoes — F3 or greater on the Fujita scale — are becoming less frequent. There is no trend, neither up nor down, in the frequency of all tornadoes.
Michael Smith is a meteorologist, the senior vice president of AccuWeather Enterprise Solutions and the author of When the Sirens Were Silent and Warnings: The True Story of How Science Tamed the Weather.