Tornadoes are not becoming more frequent overall, but when—and how many—will strike is becoming more variable.

Tornadoes are not becoming more frequent overall, but when—and how many—will strike is becoming more variable.

Minerva Studio/Shutterstock

Tornadoes striking United States in bunches

Every year, nearly 500 damaging tornadoes strike the United States, an average that hasn’t changed in recent decades. But several new studies, including one published online today in Science, are finding that tornado patterns have changed since 1970. There are more days with many tornadoes, and more days with no tornadoes at all. The start of the tornado season—typically in spring—has also become more capricious, with some years starting a month or two earlier than the norm and other years starting late.

The finding has a practical consequence: It implies that insurance companies should be keeping more capital on hand to gird themselves for large but infrequent tornado catastrophes. And emergency responders should be ready to respond to tornado clusters with sufficient equipment and manpower, says Harold Brooks, an atmospheric scientist at the National Oceanic and Atmospheric Administration (NOAA) National Severe Storms Laboratory in Norman, Oklahoma. “We need more resources … even if we don’t use them very often.”

Researchers have begun to link this increased volatility to variability in atmospheric factors that drive tornadoes, such as temperature, moisture content, and shear winds. Pointing a finger at climate change, however, is still not possible. “The links in the chain connecting them aren’t complete yet,” says Brooks, who led the Science study. Among other things, Brooks found that in the 1970s, there were on average 150 days a year with at least one damaging tornado, and that number has now dropped to 100 days. But the bad days have gotten worse: In recent years, there were on average 3 days with more than 30 tornadoes, compared with just 1 day a year 4 decades ago. In August, a study published in Climate Dynamics, led by James Elsner, a climatologist at Florida State University in Tallahassee, came to the same conclusion. “They are coming in bunches,” Elsner says.

To identify the trend, both teams had to work with a tricky database. In 1953, NOAA’s National Weather Service began collecting tornado data based on eyewitness reports, and in the mid-1970s, the agency began classifying the tornadoes based on how much damage they cause. The scheme, now called the Enhanced Fujita scale, ranks tornadoes from a 0 to a 5, from least damaging to most. But the database relies on human judgment, and the zeroes, in particular, were problematic. There appear to be many more zeroes now than there were a half century ago, simply because these relatively minor tornadoes weren’t reported as often in the past. Brooks’s and Elsner’s groups each found that by discarding these tornados and focusing on the stronger, more reliably reported ones, they were left with a relatively unbiased data set—one that put the rising variability on vivid display.

Connecting the variability to a physical cause has proved more challenging, as is predicting how tornado patterns may continue to change in a warming world. In part, that’s because tornadoes are localized phenomena that are too small to be resolved by climate models. But scientists know that tornadoes, like severe thunderstorms, are driven primarily by two atmospheric factors: a convective component associated with updrafts of warm, moist, energy-laden warm air, and a shear component associated with differences in horizontal winds that can lead to rotation. The convective component, which depends on moisture and temperature over land, will rise as the globe warms, climate scientists say. But the shear component is ultimately linked to the temperature difference between polar regions and midlatitudes, and this is expected to drop in the long term, because the poles are warming faster than the rest of the planet. The two factors may now be canceling each other out—and perhaps that has something to do with the fact that there are no more tornadoes on average these days than in the past, Elsner says.

On the other hand, some studies have suggested that the convective factor will dominate over the shear factor in the long run, producing more severe storms and tornadoes. “I don’t think we have a straight answer yet,” Elsner says. “It’s not clear how this is going to play out.