Climate change may be the existential threat of our lives, yet when it comes to in-your-face weather, tornadoes are in a class of their own. Fortunately, human-warmed climate isn’t making violent U.S. tornadoes any more frequent. However, climate change may be involved in some noteworthy recent shifts in the location and seasonal timing of the tornado threat.
The United States is the global epicenter of tornado formation. An average of about 1,200 U.S. twisters are observed each year, with some years bringing as few as 900 and others as many as 1,600-plus. It’s all the result of a unique geography that allows hot, dry air from the Southwest to flow atop moist, warm, unstable surface air east of the Rockies, with cold air at the jet-stream level overtopping it all. This layer cake of winds and air masses, varying with height, supports development of rotating supercell thunderstorms, the kind that produce the most long-lived and intense tornadoes. Even weaker non-supercell storms can collectively spawn hundreds of twisters each year.
The total number of U.S. tornadoes observed each year roughly doubled from the 1950s to the 1990s with the advent of more storm spotters and chasers (think “Twister”). Most of these “extra” tornadoes were on the weak side, though, as the more intense ones were already hard to miss. The boost provided by more eyes and cameras largely disappears when the count turns to only the 300 to 600 tornadoes per year rated at least EF1 on the Enhanced Fujita Scale (or F1 on the original scale) with top wind gusts of at least 86 mph, ignoring the forgettable “EF zeroes” (EF0s).
Each tornado is a localized creature, which makes it difficult to link to global climate trends. Climate change typically plays out in local fashion by way of broad regional shifts, such as depleted sea ice, warmer oceans, and drier landscapes. Sometimes these shifts are distinct enough from natural variation to signal clearly that human-caused climate change is likely involved. In contrast, tornadoes and their parent thunderstorms are brief and episodic, and they normally vary a great deal over time and space, so it’s tougher to distill long-term trends in their behavior and distinguish those from normal ups and downs.
Nevertheless, a few signals have shown up in tornado seasons over recent decades. Some may be the result of year-to-year or decade-to-decade variability; others could be related to longer-term, human-caused climate change. Here’s what scientists have been noticing.
Tornado seasons are getting more variable
Although the total yearly count of significant, EF1-or-stronger tornadoes (EF1+) hasn’t risen or fallen substantially or over a sustained period, how these tornadoes are distributed across time is another matter. The monthly variability of EF1+ tornadoes has increased since the 1970s, with a growing occurrence of both record-busy and record-calm months, according to a 2014 study.
To cite a recent example, the count of 510 tornadoes of all strengths in May 2019 was more than 100 above any other May on record. Just two years later, May 2021 became the first May on record without a single EF3-or-stronger tornado (EF3+) anywhere in the United States.
What’s causing the increased variability? No single culprit has been identified, but the researchers of that same study found that variability is growing on the annual scale also. One example: The most active and destructive tornado year in modern records, 2011, was followed by one of the quietest, 2012. More recently, the year 2018 had a record-low death toll, with only 10 tornado-related fatalities, and it was the first year with no EF3+ tornadoes reported.
Tornado outbreaks are getting larger and more frequent
Increased variability extends to the daily scale too. A 2016 study found that an increasing number of each year’s tornadoes are occurring in outbreaks (periods of one to several days with at least six closely spaced EF1+ tornadoes). The study also found that outbreaks themselves are becoming more frequent. Historically, close to 80% of all U.S. tornado deaths are associated with outbreaks.
On the flip side, this clumping of an annual tornado crop that’s not significantly growing or shrinking has left an increasing number of days free of any tornadoes of EF1+ strength.
As with monthly and annual variability, no scientific studies have shown a link between the increase in daily-scale variability and climate change, and there is no certainty it will continue into future decades. The concept does bear some similarity to the observed trend – strongly connected to climate change, and expected to continue – of a larger share of U.S. rainfall falling in more intense episodes, though the two phenomena have not been explicitly studied together.
If nothing else, the increasingly mercurial timetable of twisters suggests we ought not be too shocked if we happen to get a hyperactive tornado month or season, or an ultra-sleepy one.
Is tornado season starting earlier?
Events such as a massive “warm wave” in March 2012 that sent temperatures of 90°F as far north as Michigan raise questions of whether tornado activity might shift earlier in the year as U.S. (and for that matter global) warming continues. There’s scant research addressing this topic.
Another study published in 2014 examined the belt traditionally known as Tornado Alley, from northern Texas to southeast Nebraska, and it found that the region’s yearly springtime peak of tornado activity had shifted from around May 25 to May 14. However, a follow-up study found no such trend toward an earlier peak across a patch of the southeastern U.S. from Arkansas to northern Alabama where overall tornado activity has been increasing (see below).
It’s important to keep in mind that tornadoes can develop anywhere, at any time of year – including Wisconsin in January – as long as the proper ingredients are in place.
Tornado Alley is shifting eastward
Perhaps the most concerning trend in recent decades is geographic. A 2018 study found that over the past 40 years, EF1+ tornadoes have increased in frequency from roughly Louisiana to Missouri eastward, especially south of the Ohio River, east of the Mississippi, and west of the Appalachians. Many of the deadliest and most destructive tornadoes of the 21st century have occurred in that particular region, including those in the catastrophic Super Outbreak of 2011 as well as more recent disasters such as the Tennessee tornadoes of 2020 that caused billions in damage and killed 28.
Looking at weather conditions that support tornadoes, rather than tornadoes themselves, the same study found a similar eastward shift.
At the same time, EF1+ tornadoes and tornado-favorable conditions have become less frequent over most of the traditional Tornado Alley states from Texas to Nebraska. (But the annual average tornado count remains higher in those states than further east.) Tornado activity hit unprecedented lows across large parts of Oklahoma and Kansas from 2020 into 2021.
A USA TODAY analysis in June 2021 vividly brought home the location shift.
Like the other trends outlined above, the eastward shift in tornadoes has not yet been conclusively linked to any particular aspect of climate change, and it’s uncertain whether this geographic shift will continue into future decades. One factor that could be involved is strong multidecadal warming observed across the U.S. Southwest. The increased heat may be generating air just above the surface that’s hot enough to more reliably suppress tornadic thunderstorm development as the air flows eastward above traditional Tornado Alley. Meanwhile, sea surface temperatures across the Gulf of Mexico are increasing, which helps to generate warm, humid surface air feeding into severe thunderstorms across the central and eastern U.S. How these factors and others may be influencing the eastward shift in tornado prevalence remains to be determined.
Tornado vulnerability is on the rise
The eastward shift is especially ominous because of the Southeast’s vulnerability on multiple counts. More people live in tornado-vulnerable manufactured homes in the Southeast than anywhere else, and many of these are on small acreages where safe shelter can be miles away. Southeastern tornadoes are also more likely than those elsewhere to strike at night, when many people are asleep and approaching tornadoes are less visible.
Increasing population will add to the nation’s tornado vulnerability going forward, even without considering any climate-related changes to tornado behavior. A 2017 analysis found that average yearly tornado impacts and vulnerability could be 6 to 36 times higher by 2100 compared to 1940, depending on location. The biggest projected increase is in the Mid-South region from eastern Arkansas to the Appalachians.
Even locations such as New York City are at a surprisingly high tornado risk when considering the extreme impacts that a low-probability strike could inflict on a dense, vulnerable urban population, as documented in the National Risk Index released by FEMA in 2021.
The tornado outlook for later this century
Only a few studies have attempted to simulate how tornadic thunderstorms might behave in the expected warmer climate of the late 21st century. One approach is to examine tornado environments – the weather configurations that support tornadic thunderstorms – because these are more readily simulated in a climate model than tornadoes themselves.
Initial work in this area found that the instability fueling severe thunderstorms will likely increase over the century, but the vertical wind shear needed for tornadic supercells will more likely decrease. The projected result would be an increase in overall severe weather (including heavy rain and gusty wind), but a potential drop in tornado frequency.
Subsequent research, looking at how the juxtaposition of instability and wind shear might itself change, came up with different results. In a nutshell, the springtime days with enhanced instability tended to also have tornado-supportive wind shear, implying there could be an increase in tornadoes after all.
Some model configurations can now go a step further, “downscaling” the global model output to see how actual thunderstorms might evolve in the projected future climate. So far, the results tend to confirm earlier research that severe weather overall will become more frequent, perhaps with a longer storm season. At least one study suggests that variability in peak-season behavior will increase, in line with recent trends.
Even in these high-resolution models, tornadoes themselves are far too small to be directly simulated, so that reality remains a major caveat.
One other persistent challenge is simulating and analyzing the weather features that can inhibit tornadic storms on even favorable-looking days, such as the “caps” of warm, dry air one to two miles above the surface. Some of the 21st-century modeling suggests that summertime capping may intensify later in the century.
Harold E. Brooks, Gregory W. Carbin, and Patrick T. Marsh, 2014: Increased variability of tornado occurrence in the United States. Science, 346, 349–352.
John A. Long and Paul C. Stoy, 2014: Peak tornado activity is occurring earlier in the heart of “Tornado Alley”. Geophysical Research Letters, 41, 6259–6264.
John A. Long, Paul C. Stoy, and Tobias Gerken, 2018: Tornado seasonality in the southeastern United States. Weather and Climate Extremes, 20, 81–91.
Vittorio A. Gensini and Harold E. Brooks, 2018: Spatial trends in United States tornado frequency. Climate and Atmospheric Science.
Michael K. Tippett, Chiara Lepore, and Joel E. Cohen, 2016: More tornadoes in the most extreme U.S. tornado outbreaks. Science, 354, 1419–1423.
Stephen M. Strader, et al., 2017: Observed and Projected Changes in United States Tornado Exposure. Weather, Climate, and Society, 9, 109–123.
Robert J. Trapp, et al., 2007: Changes in severe thunderstorm environment frequency during the 21st century caused by anthropogenically enhanced global radiative forcing, Proceedings of the National Academy of Sciences (PNAS), 50, 19719–19723.
Noah S. Diffenbaugh, Martin Scherer, and Robert J. Trapp, 2013: Robust increases in severe thunderstorm environments in response to greenhouse forcing. PNAS, 110, 16361–16366.
Kimberly A. Hoogewind, Michael E. Baldwin, and Robert J. Trapp, 2017: The Impact of Climate Change on Hazardous Convective Weather in the United States: Insight from High-Resolution Dynamical Downscaling. Journal of Climate, 30, 10081–10100.