Are extreme swings in tornado activity the new normal in the US?

The likelihood of a single tornado forming on any given day is falling. However, the country is likely to face more outbreaks featuring 20 to 30 or more twisters in a single day if recent trends in tornado activity in the United States hold up.

That's the implication of a new analysis of US tornado records by a trio of researchers and forecasters at the National Oceanic and Atmospheric Administration's National Severe Storms Laboratory and the Storm Prediction Center, both in Norman, Okla.

The researchers found that between 1954 and 2013, the average number of tornadoes that ​touch down in the US each year show no statistically significant change when the weakest category of tornado intensity was excluded. But during the past 15 years, the number of months with extreme activity – either high or low – increased, with a virtually even split between months with extremely high and extremely low activity.

“What you end up with is a boom or bust” in extreme tornado activity, even as the average number of twisters a year remains the same, says Harold Brooks, a researcher at the National Severe Storms Laboratory and the lead author of the analysis, which appears in ​Thursday's issue of the journal Science.

​The increased frequency of these tornado extremes has occurred even as the total number of days with any tornado at all has declined, and the length of the tornado season – a fiendishly difficult concept to establish, since tornadoes can occur in every month of the year – has grown. For this study, the team defined the season as starting on the day producing the fiftieth EF1 tornado, the second weakest category, which ranges from EF0 to EF5.

The extremes appear to be growing more common by another measure, he adds. From the 1950s through the 1970s, only one year saw two days where severe weather spawned more than 30 tornadoes, Dr. Brooks says. Since 2002, each year has seen at least two days with 30 tornadoes. By that measure, what once was a rare event has become common, he says.

The results mesh with those of another analysis of tornado trends conducted by researchers from Florida State University in Tallahassee and published in August in the journal Climate Dynamics. The team, led by James Elsner, whose research has focused largely on tropical cyclones, found a similar increase in tornado extremes between 1954 and 2013.

They looked at changes in the number of days with many tornadoes each year relative to the number of days with at least one tornado. They found that the number of days with many tornadoes was increasing while the number of days with any tornadoes was decreasing.

Moreover, they parsed the data in ways that they say reveals an increase in the number of tornadoes clustered in the same general area.

Dr. Elsner's team noted that their results were “broadly consistent” with modeling studies that project an increase in conditions that favor the formation of severe thunderstorms with global warming – although the proportion of severe storms that spawn tornadoes is small and the factors that trigger tornado formation beneath any given thunderstorm, particularly conditions within the first 600 feet above the ground, are poorly understood.

For its part, Brooks's team says it can offer no specific physical explanation for the increasing variability in extremes and the increase in their frequency at this point. But the researchers also acknowledge that the answer probably involves either changes to the general atmospheric environment or the more-limited environment in which a funnel cloud develops and drops to form a tornado.

Whatever the cause, if the trend continues, it has significant implications for emergency managers at the state and federal levels, who will need to have sufficient supplies and sufficient boots ready to hit the ground to deal with increases in tornado outbreaks. The trend also could force insurers and reinsurers to alter their rate structures to deal with a boom-and-bust pattern of tornado extremes.

In the meantime, taking a cue from the development of seasonal hurricane forecasts, researchers are exploring ways to producing seasonal tornado forecasts as a way to help emergency managers and insurance companies gauge their needs and risks for upcoming tornado seasons.