NOAA Storm Prediction Center · Raw severe-weather database · Tornadoes 1950–2025 · Wind 1955–2025

Are severe storms in Vermont, New Hampshire & Maine getting more frequent or stronger?

Two questions, answered from NOAA's raw event record for both tornadoes and severe thunderstorm / high-wind events: how often severe storms strike (frequency) and how hard they hit (strength).

✅

Short answer: No clear increase in either, with a cautionary lesson hiding in the wind data.

Tornado frequency is flat-to-slightly-down. Severe wind reports rose about 57× since the 1950s, but that is overwhelmingly a reporting artifact (more spotters, more people and changed NWS practices), not more storms. Over the very same period, measured wind strength fell. Neither hazard shows any drift toward stronger events either: the share of violent tornadoes (EF2+) and significant wind gusts (≥75 mph) both declined. Frequency of reports ≠ frequency of storms ≠ intensity of storms.

Tiles marked in orange reflect report-count changes that are driven largely by how storms are observed and logged, not by the weather itself.

How OFTEN

frequency of events

Counting events per year. Tornadoes are individually surveyed and reasonably consistent over time; raw wind reports are not, so read them with the caveat below.

Show tornado data table

Tornadoes per year, by state (1950–2025)

Stacked bars split each year by state. Solid line = 5-year moving average; dashed line = linear trend across all 76 years, sloping gently downward.

Vermont New Hampshire Maine 5-yr moving avg Linear trend

Severe thunderstorm-wind reports per year (1955–2025)

Every logged severe wind event (gusts ≥ 58 mph / 50 kt) across the three states. The apparent explosion is the classic severe-weather data pitfall.

Wind reports (VT+NH+ME) 5-yr moving avg

⚠️ This is not the weather changing. It's the record changing. Reports climbed from ~5/yr in the 1950s to ~288/yr in the 2010s (≈57×). The jumps line up with more trained spotters, rising population and NWS moving to storm-based verification, not with any physical surge in storms. The strength charts below show why: as reports multiplied, the average reported gust got weaker, because the database filled with marginal, threshold-level events.

Tornadoes: average per year, by decade

A mid-century peak, a long dip, a modest 2010s rebound, then a fresh decline. (2020s = 2020–2025.)

Wind reports: average per year, by decade

A near-monotonic climb that mirrors reporting history, not climate. Shown only to make the artifact legible.

How HARD

strength / intensity of events

Frequency and intensity are separate questions. Here we set counts aside and ask whether the storms that do occur are getting more violent. For tornadoes: the EF/F damage rating. For wind: the recorded gust speed. Neither is trending up.

Tornadoes: share rated EF2+ (strong / violent)

Of tornadoes with a damage rating, the fraction reaching EF2 or higher, by decade. It falls to zero in the 2010s–2020s, when recent tornadoes here are overwhelmingly weak (EF0–EF1).

Tornadoes: average EF rating by decade

Mean damage rating of rated tornadoes. Same story from a different angle: intensity drifts down, from ~1.3 in the 1960s to ~0.4–0.6 recently.

Wind: gust strength by decade (mph)

Median (typical) and 90th-percentile (high-end) recorded gust. The typical gust is flat near the 58 mph reporting floor; the strongest gusts declined. Shown from the 1980s on, once speeds were recorded often enough to be reliable.

Median gust 90th-percentile gust

Wind: share of significant gusts (≥ 75 mph)

Fraction of wind reports reaching the "significant severe" threshold, by decade. It collapses from ~19% to ~2%, again the signature of an expanding, marginal-heavy report set rather than weakening physics alone.

Each state on its own: tornadoes per year, by decade

No state shows a sustained climb. Maine (most active) peaked in the 1970s and 2010s; all three fall off in the 2020s.

Vermont New Hampshire Maine

How to read this and its limits

Report counts are not storm counts, especially for wind. Severe-wind reporting inflation is the best-documented bias in this dataset. The ~57× rise reflects more spotters, higher population and the shift to storm-based verification. Take the raw wind-frequency trend as a lesson in data hygiene, not evidence of a stormier climate.
The falling strength numbers are partly the same artifact in reverse. As reporting widened, both databases filled with marginal events (EF0 tornadoes; 58 mph threshold gusts) that earlier decades never recorded. That mechanically pulls the average intensity and the "significant share" down. The honest reading: there is no evidence of strengthening, but the apparent weakening is largely a recording effect, not necessarily physics.
Tornadoes are the cleaner signal. Each is individually surveyed and damage-rated, so tornado counts and EF ratings are far less biased than raw wind reports. They show flat-to-down frequency and intensity.
Small numbers, big noise. These three states see only a handful of tornadoes a year; single outbreak days swing an entire year. Only multi-decade shapes are meaningful. Pre-1980 wind speeds were recorded too rarely to trend reliably, so the wind-strength charts start in the 1980s.
Scope: tornadoes and thunderstorm/high wind only, not hail, rainfall or flooding, which can trend differently. Wind gust ≥ 58 mph (50 kt) is the severe threshold; ≥ 75 mph (65 kt) is "significant severe."

Source: NOAA / NWS Storm Prediction Center Severe Weather Database (WCM): the 1950–2025 Actual Tornadoes and 1955–2025 severe-wind period files, spc.noaa.gov/wcm. Filtered to state = VT, NH, ME (7,561 wind reports; 282 tornadoes). Wind speed converted knots→mph (×1.15). Tornado trend = OLS on annual counts (slope −0.032/yr, r = −0.22). Raw CSVs and computed series saved alongside this file. Generated 2026-07-14.