Tornado Statistics: The 2025 U.S. Snapshot
Published September 15, 2026 · StatsGalore
Preliminary tornado count in NOAA’s annual assessment [1]
NOAA’s January 2026 annual assessment listed a preliminary U.S. tornado count of 1,559 for 2025. The report placed it at 127% of the 1991–2020 average. These are dated preliminary figures and can differ from later finalized storm records. [1]
The numbers
| Measure | Reported figure | Source |
|---|---|---|
| Preliminary annual count | 1,559 | [1] |
| Annual count as a share of 1991–2020 average | 127% | [1] |
| Preliminary March reports | 300 | [1] |
| North Dakota annual reports | 72 | [1] |
Preliminary means subject to revision
Early storm reporting can contain duplicate reports or events that need further investigation. Later assessments can confirm, combine, or revise records. The reporting date and preliminary status should therefore remain attached to the count.
A more recent finalized database may show a different number without either source having been dishonestly reported. The difference can reflect the stage of verification.
[1]Reports and confirmed events need clear labels
The annual assessment uses preliminary counts and reports in different parts of its discussion. The table keeps those qualifiers instead of treating every row as a finalized count under an identical method.
The March and North Dakota figures are subsets by time or location, not extra categories to add to the annual national total.
[1]A share of an average is not a percentage increase
A count equal to 127% of a reference average is above that average; it does not mean an increase of 127%. The reference period also matters because a different comparison window can produce a different ratio.
Historical comparisons require care about changes in observation and reporting. A national count alone is not a complete measure of changes in storm intensity or exposure.
[1]Frequency is different from personal risk
The national event count does not give the probability that a particular home will be struck. Location, season, storm paths, and exposure affect that question. Dividing national population by tornado count would not produce a meaningful personal risk estimate.
This page is a historical statistical snapshot. It does not provide current warnings or establish whether weather conditions are safe at a particular time.
[1]Sources & notes
NOAA NCEI annual assessment published January 13, 2026. Figures explicitly retain preliminary/report qualifiers; later confirmed datasets may revise them.
Source review: 2026-09-15. Article publication and underlying data dates are separate. How we use statistics · Suggest a correction