Data analysis
Tornado fatality statistics
Tornado deaths are not random. Certain locations, structures, ages, and situations produce disproportionate fatalities. Here is the data.
The big picture
- US average: 71 tornado deaths per year (1990-2020).
- Range: 15 to 553 depending on year.
- Worst year: 2011 (553 dead, mostly Super Outbreak).
- Best year: 1988 (32 dead).
- Fatality rate per tornado has decreased over decades — thanks to warnings.
- Total tornadoes has increased due to better detection.
Fatalities by structure
Mobile / manufactured home
43% of deaths despite <8% of housing. 15-20x death rate of site-built.
Vehicles
11% of deaths. Higher risk overnight and rural.
Outdoors
10% of deaths. Often people who tried to observe the tornado.
Site-built home
32% of deaths.
Business / commercial
4% of deaths.
Community shelter
<1% of deaths.
Fatalities by time of day
- Nighttime (7 PM - 7 AM): ~40% of tornadoes, but ~65% of fatalities.
- Peak fatal hour: 9-11 PM.
- Dawn (5-7 AM): 15% of fatalities.
- Daylight tornadoes: relatively fewer fatalities per warned event.
- People asleep or without visual cues die at 2.5x the rate.
Fatalities by region
Dixie Alley (MS/AL/AR/TN)
40% of all US tornado deaths despite <20% of tornadoes.
Great Plains (KS/OK/NE/TX Panhandle)
25% of deaths.
Southeast (GA/SC/NC/FL)
15% of deaths.
Midwest (IL/IN/MO/IA)
15% of deaths.
Northeast
<5% of deaths.
West of Rockies
<1% of deaths.
Fatalities by age
- Under 5: elevated risk (relies on adults for protection).
- 65+: highest death rate (mobility, isolation, older housing).
- 75+: 3x death rate of adults 25-64.
- Working-age adults: lower death rate but higher raw count.
- Children in vehicles or mobile homes: extreme risk.
Fatalities by tornado intensity
EF0
<1% of tornado deaths — but many injuries.
EF1
5% of deaths.
EF2
10% of deaths.
EF3
25% of deaths.
EF4
35% of deaths.
EF5
25% of deaths.
EF4-EF5 tornadoes are rare (<1% of all tornadoes) but produce 60% of fatalities. This is why targeting these events matters.
The deadliest tornadoes in US history
- 1925 Tri-State — 695 dead. Deadliest ever.
- 1840 Natchez, MS — 317 dead.
- 1896 St. Louis — 255 dead.
- 1936 Tupelo, MS — 216 dead.
- 1936 Gainesville, GA — 203 dead.
- 1947 Woodward, OK — 181 dead.
- 1908 Amite, LA — 143 dead.
- 1899 New Richmond, WI — 117 dead.
- 2011 Joplin, MO — 158 dead. Deadliest US tornado since 1947.
- 1953 Flint, MI — 116 dead.
- 1953 Waco, TX — 114 dead.
- 1902 Goliad, TX — 114 dead.
What has changed since 1875
- 1875-1900: No warnings existed. Fatalities were catastrophic.
- 1900-1950: Weather Bureau prohibited tornado warnings — feared panic. Deaths accumulated.
- 1950s: Public warnings began. Deaths dropped modestly.
- 1970s: Radar era. Warning lead time grew from 0 to 5 min.
- 1980s-1990s: Doppler radar. Lead time 11-13 min. Deaths dropped further.
- 2000s: Cell alerts began. WEA in 2012.
- 2010s-2020s: Impact-based warnings. Some recent years buck the trend (2011, 2021).
Preventable deaths
Most tornado fatalities in the modern era are preventable. Common patterns:
- Person in mobile home who didn't have alternate shelter.
- Person in vehicle trying to outrun tornado.
- Person outside recording video.
- Person didn't receive warning (WEA off, sleeping, hearing impaired).
- Person believed the warning was false (warning fatigue).
- Person believed the tornado wouldn't hit their specific location.
- Person had no shelter plan.
- Person's "shelter" was actually a poor choice (attic, bathroom on 2nd floor).
The equity dimension
- Mobile home fatality rate 15-20x site-built homes.
- Mobile home density is higher in low-income and rural areas.
- Non-native English speakers face warning language barriers.
- Overnight-shift workers face different warning exposure.
- Historically, warnings were only in English.
- NWS has expanded Spanish-language warnings.
- Community shelter access disparities.
- Insurance coverage disparities post-storm.
What continues to save lives
- NOAA Weather Radio with SAME.
- Wireless Emergency Alerts (WEA).
- Storm shelters and safe rooms.
- Skywarn spotter network.
- Broadcast meteorology continuity.
- Impact-based warnings.
- Community shelter programs.
- Family communication plans.