Atmospheric conditions capable of spawning major tornado outbreaks across the United States could expand significantly into the Midwest, Great Lakes and Northeast between 2065 and 2099, according to new climate research.
The projected shift specifically applies to May, historically the peak month for major US twisters, driven by a warmer and wetter atmosphere combined with changing jet stream and wind patterns.
Researchers warned that traditional tornado hazard zones will not necessarily become safer as the atmospheric threat footprint spreads across a broader portion of the nation.
Historically, Tornado Alley has stretched through the central Great Plains, encompassing states such as Texas, Oklahoma, Kansas, Nebraska and South Dakota, while Dixie Alley has covered southeastern states including Mississippi, Alabama and Tennessee. Under the new climate projections, outbreak-supporting conditions could become more frequent in Missouri, Illinois, Indiana, Iowa, Minnesota and Wisconsin, reaching as far east as Pennsylvania, Virginia and New York.
Dr Jana Houser, an associate professor of meteorology in the atmospheric sciences program at The Ohio State University who was not involved in the research, urged communities across the eastern half of the country to prepare. She told Daily Mail that the entire region needs to have conversations about what increased tornado potential means for families, stressing that everyone should have safety plans in place because it only takes one tornado to change lives.
Houser cautioned that the study tracks changes in weather environments that support severe storms, rather than predicting the exact number of tornadoes each region will experience, adding that the central Plains could still record the highest total number of tornadoes nationwide.
Climate model projections and emissions pathways
The study, published in the peer-reviewed journal npj Climate and Atmospheric Science, was conducted by a multi-institutional research team from the University of Oklahoma, the Massachusetts Institute of Technology, the National Oceanic and Atmospheric Administration, and NASA.
To project future trends, the team analyzed atmospheric patterns surrounding 45 major May tornado outbreaks recorded between 1980 and 2014, then tested that environmental fingerprint in a high-resolution global climate model across four distinct greenhouse gas emissions pathways over 35-year periods.
Under an intermediate emissions pathway, favorable outbreak conditions expanded from eastern Texas and Oklahoma through the Mississippi and Tennessee valleys to Virginia, Pennsylvania and New York. Higher emissions shifted the primary core northeastward, producing significant increases in Tennessee, Kentucky, southern Illinois and southern Indiana.
An extreme warming scenario generated the widest overall geographical footprint, yielding the largest gains in Wisconsin, Minnesota, Iowa and Illinois, with the strongest signal appearing in eastern Missouri. Conversely, western Florida showed an opposite trend, recording a projected decline in outbreak-supporting conditions.
Lead researcher Paulina Cwik of the University of Oklahoma told Daily Mail that what stood out to her was how projected patterns spread farther north and east while remaining present in areas already facing high major outbreak risks. She noted that rather than seeing one tornado-prone region simply replaced by another, the atmospheric patterns associated with major outbreaks could extend across a broader geographical area.
Atmospheric ingredients and proxy day variations
Meteorologists explain that tornado formation relies on two critical atmospheric ingredients: abundant moisture and wind shear, which is the variation in wind speed and direction with altitude that creates rotation inside severe thunderstorms. Warmer air holds more moisture, while shifts in the jet stream and the Great Plains low-level jet can redirect that moisture and alter wind shear.
However, extreme warming could eventually weaken some of those ingredients by reducing midlatitude wind shear and strengthening the atmospheric cap, a layer of warm air aloft that acts as a lid to prevent thunderstorms from forming.
This complex relationship explains why the model identified 80 historical outbreak-proxy days over a 35-year baseline, rising to 85 days under the lowest-emissions pathway, 100 days under the intermediate pathway, and 112 days under the high pathway, before falling to 93 days under the most extreme scenario.
Cwik expressed surprise that the relationship with future climate scenarios was not simple, stating, "The highest-emissions scenario we examined did not produce the largest number of outbreak-supportive days. Instead, the results varied across scenarios, both in the number of outbreak-supportive days and in how the associated atmospheric patterns were organized geographically."
In annual terms, the totals represent an increase from 2.29 outbreak-supporting days each May historically to between 2.39 and three days in future simulations. Researchers emphasized that this overall rise was not statistically significant because tornado-supporting weather varies dramatically between years, making the spatial redistribution of favorable conditions a more reliable finding than any increase in outbreak frequency.
Houser noted that there is substantial year-to-year variability in outbreak numbers and locations, adding that one year might see very low numbers while another sees very high totals in different areas.
Expanding geographic footprint and model limitations
The study found that the geographical area exposed on each proxy day expanded from roughly 328,000 square miles historically to about 386,000 square miles under the low-emissions pathway and 402,000 square miles under the intermediate scenario, representing an increase of up to 22 percent.
Under the most extreme pathway, the portion of the study area exceeding one high-end atmospheric threshold rose from 3.3 percent to 8.1 percent, a 146 percent increase that Cwik explained points to a reorganization of broader atmospheric patterns rather than proof that individual outbreaks will cover more territory.
Houser warned that a larger environmental footprint could put more people at risk, but stressed that global climate models cannot resolve small-scale physical details like land cover, terrain, and localized storm dynamics that determine whether a tornado forms. She explained that out of six storms in an identical supportive environment, only two might produce tornadoes for reasons meteorologists do not yet fully understand.
Cwik emphasized that the analysis does not allow researchers to say that a future tornado outbreak will necessarily cover a larger area, produce more tornadoes, or expose a specific number of additional people, which would require storm-resolving simulations together with population and exposure analyses. She noted that because global climate models cannot explicitly simulate individual tornadoes or storm-scale processes such as low-level rotation, the results represent changes in supportive weather patterns rather than direct projections of tornado occurrence or intensity.
The researchers stressed that the findings come from a single climate model, examine only May, and rely on fixed atmospheric thresholds that may behave differently in a warmer world. Cwik described the results as a proof of concept rather than a settled forecast, while Houser called for future studies across every month using multiple model configurations to build scientific consensus on future tornado behavior.
