How many people did tornadoes kill in the US in 2025, how many of those deaths came while a tornado watch was in effect, and what did storm chaser Joshua Wurman, founder of the Doppler on Wheels radar network, say about chasing tornadoes and about warnings when Book of Odds spoke with him in 2009?
In 2025, 23 tornadoes killed 67 people in the US, final data from the NOAA Storm Prediction Center (SPC) show. SPC’s preliminary list, issued in January 2026, put 61 of its 67 deaths in tornadoes inside a tornado watch; the final database does not record watches. To see how tornado deaths compare with heat and floods, read our full comparison.
Book of Odds spoke with Joshua Wurman by phone on September 29, 2009.
Key numbers
| Figure | Source and year |
|---|---|
| 23 killer tornadoes | NOAA SPC, 2025, final |
| 67 tornado deaths | NOAA SPC, 2025, final |
| 61 of 67 deaths in a tornado watch | NOAA SPC, 2025, preliminary list |
| 309 to 318 mph calculated peak wind, Greenfield, Iowa | FARM, May 21, 2024 tornado, preliminary |
| 15 tornado deaths, 2026 through June 23 | NOAA SPC, 2026, preliminary |
The interview
What follows is the original 2009 Book of Odds feature, with the kept quotes unchanged. Its counts are as of 2009: the “nearly 15 years” of chasing, the “more than 140 tornadoes” and “last spring,” which means spring 2009. The figures in his quotes are his, from 2009, and were not checked by us. The original opening lines, which included one quote, are cut, and one paragraph is replaced with newer data.
Original Book of Odds feature:
Wurman is so confident he will survive an up-close encounter with a tornado, he has spent nearly 15 years chasing them all over the world. In his opinion, “People dread tornadoes out of proportion to the actual probability that they will happen.”
Wurman is the founder of the Doppler on Wheels (DOW) radar network. On May 21, 2024, a DOW radar measured Doppler velocities as high as 263 to 271 mph in a tornado at Greenfield, Iowa, and DOW scientists calculated peak winds of 309 to 318 mph, among the highest ever determined from DOW data, according to a statement from the Flexible Array of Radars and Mesonets (FARM). FARM says these values are preliminary. As of 2009, his team had been featured on the Discovery Channel’s Storm Chasers, but even Wurman can be fooled by Mother Nature.
Original Book of Odds feature:
“Most storms we intercept don’t turn into tornadoes. And we really don’t have great skill in knowing which ones are going to, so we chase after a lot of them. Only a few pay off.”
To date, Wurman has collected data from more than 140 tornadoes. “The prize is not only getting the data, it’s learning from that data. We’re trying to understand as much as we can, not only about the wind speed, but also the temperatures and the relative humidity in and around tornadoes.”
Original Book of Odds feature:
Most of the year he spends in the lab. But when the peak tornado months of May and June come around, Wurman’s team sets out. Last spring, Wurman spent five weeks chasing storms as one of the leading scientists involved with the largest tornado research project to date: VORTEX2.
Tornado Season 2009 turned out to have many quirks. “We were kinda shut out in the first three weeks. There were hardly any storms, never mind tornadoes.”
They observed only one tornado. But what a tornado it was. “That particular tornado was by far the best-studied tornado ever. That was a platinum-plated, diamond-encrusted data set! The tornado lasted 30 minutes, and we’ll be looking at that data for 30 months.”
When the spring storms finally kicked up, the teams had to constantly adjust their pace. “Storm chasing is a lot of shifting, jockeying for positions. We start out the morning in a hotel somewhere, waking up around 10 AM. We’d do weather briefings and discuss where we think the target area’s going to be. Sometimes it was, ‘Oops, you know, the storm’s really going to form out in Colorado, and we have to drive six hours to get there!’ We’d scramble to leave right away.”
Frantic activity is interspersed with long stretches of downtime with little to do. “We’d be in a Wal-Mart parking lot for two or three hours, or maybe a gas station.” Waiting for a storm to take shape, “We like to be where there’s at least some junk food and a bathroom.” With his large crew, there is always the issue of logistics: getting everyone fed, and the vehicles fueled. “Imagine 160 hungry people showing up in one of these small, one-gas station towns.”
When they think a storm is about to “pop,” they make a quick decision to try and intercept the storm. But it’s easy to get lost.
“Every minute we’re watching the tornado change and looking at it with multiple radars. We’re surrounding the tornado with weather instruments mounted on cars and on tornado pods that we drop in front of the tornado.”
“Our entropy increases with time. We just get more random in our deployment.”
They hope that the data they collect will help develop better forecasting strategies. As it stands now, Wurman identifies three major issues with tornado prediction.
Timing: “Warnings don’t have enough lead-time. The average lead-time is 13 minutes, which means that half of those warnings are happening with less than 13 minutes. What’d we’d like is to give people a half hour to an hour of lead time ahead of the storm.
Accuracy: “There’s 70% false alarm rate associated with warnings. So do you really head to your basement? Do you really wake up your kids and have everybody screaming and ruining your evening?”
Precision: “It isn’t an official National Weather Service goal, but I think we need to get warnings more precise, telling us whether or not a really violent tornado is coming our way.” He compares it to hurricane warnings. “When a hurricane’s coming and you’re told it’s a 1, put some plywood on and stay home. If it’s a category 5, everybody below 30 feet in the coastal areas better evacuate.”
The short version
Final SPC data show 23 killer tornadoes and 67 tornado deaths in the US in 2025. SPC’s preliminary list put 61 of its 67 deaths in tornadoes inside a tornado watch.
Last checked: 2026-10-22, against the NOAA Storm Prediction Center’s final 2025 tornado database, its preliminary killer tornado lists for 2025 and 2026 and FARM’s Greenfield, Iowa statement. This article was machine-checked against the sources above and has not been expert-reviewed. Spotted an error or a newer public figure? Please tell us on our contact page with the article link and we will review it.
Sources
- S1 NOAA NWS Storm Prediction Center. Severe Weather Database Files, 2025_torn.csv, final 2025 tornado data (updated April 23, 2026). https://www.spc.noaa.gov/wcm/ Accessed Oct 22, 2026. Also 2025 Preliminary Killer Tornadoes (issued January 12, 2026), for the tornado watch count. https://www.spc.noaa.gov/climo/torn/STATIJ25.txt Accessed Oct 22, 2026.
- S2 NOAA NWS Storm Prediction Center. 2026 Preliminary Killer Tornadoes (issued June 23, 2026). https://www.spc.noaa.gov/climo/torn/STATIJ.txt Accessed Oct 22, 2026.
- S3 Flexible Array of Radars and Mesonets (FARM). FARM statement concerning preliminary peak wind speed determinations based on Doppler On Wheels (DOW) data obtained in the Greenfield, Iowa tornado of 21 May 2024. June 23, 2024. http://www.farmfacility.org/pdf/best-greenfield-windspeed-note-2024-0623bp2.pdf Accessed Oct 22, 2026.
- S4 Book of Odds. Phone interview with Joshua Wurman, September 29, 2009 (original feature).