Funnel Cloud in the United States

How dangerous are Funnel Cloud events in the U.S.?

The annual record is uneven rather than steadily rising or falling. The peak year on record was 2015, logging 362 events and $0 in property damage in a single calendar year. The most recent year in the dataset is 2025 with 185 events recorded. Across the full 11-year window, annual event counts reflect natural year-to-year noise given how infrequently this hazard is recorded at all.

Where this hazard actually strikes varies sharply by region. The three states absorbing the largest share of events are Iowa (374 events, 0 fatalities), Florida (270 events, 0 fatalities), North Dakota (255 events, 0 fatalities), a pattern driven more by climatology than by any one state's vulnerability. At the other end, states such as New Hampshire, Vermont, American Samoa report far fewer events, often reflecting hazard climatology rather than underreporting.

This is a more frequency-heavy than loss-heavy record. The hazard ranks #25 by event count but #48 for combined property and crop damage, a gap of 23 places. Its listed losses represent <0.1% of the catalog total. The human-toll rank is #48, or 0% of recorded fatalities, so neither event count nor damage rank substitutes for the other.

This hazard is recorded far more often than it proves fatal. Among the 44 hazard types with at least 100 recorded events it ranks #25 by frequency but only #43 by fatalities per event, at no recorded fatalities. Recorded damage averages $0 per event, so its cumulative toll is driven mainly by how routinely it occurs rather than by the severity of any one event.

The cumulative NOAA record has a smaller reported human toll than most hazard types on this measure: 0 fatalities and 2 injuries among 2,930 events. PlainHazard's low human-toll band means fewer than five recorded fatalities during 2015-2025. It does not mean that an individual event is harmless, or that unrecorded impacts are zero.

NOAA entries are reports filed by local National Weather Service offices. Counts therefore reflect both observed weather and reporting practice; they do not measure unreported events, household exposure, or the probability of a future event at one address.