Findings: Halloween is usually dry, and recent Halloweens ran cooler, not warmer
Original analysis by Holiday Weather Odds of 31 Halloweens (1995โ2025) at 310 NOAA stations. Free to cite with a link back.
74%of Halloweens were dry in the median U.S. city
-2.5ยฐFmedian change in a city's typical Halloween high, 2010โ25 vs 1995โ2009
92 of 306cities had a warmer typical Halloween high in the later period
2013wettest Halloween nationally (59% of stations wet)
What the data show
- Rain on Halloween is the exception. The median city had precipitation on only 26% of Halloweens. See every city.
- Recent Halloweens ran cooler, against the warming trend you might expect. Comparing each city's median Halloween high in 1995โ2009 to 2010โ2025, the median change is -2.5ยฐF, and only 92 of 306 cities (30%) came out warmer. This is one calendar day sampled 15โ16 times per period, so a few cold Halloweens (and where they fell) drive it. It says nothing about climate trends, which need all days of the year.
- Thanksgiving moved -0.1ยฐF by the same comparison (median across cities). Thanksgiving by city.
- Bad Halloweens are national events. The wettest Halloweens across all stations were 2013 (59% of stations wet), 2019 (55% of stations wet), 1995 (54% of stations wet); the driest were 2007 (10%), 2008 (7%), 2020 (5%).
- Best trick-or-treat odds: Prescott, AZ was dry with a 50โ80ยฐF high on 90% of Halloweens. Snow on Halloween is rare but real in Fairbanks, AK (35%), Northway, AK (33%), Mount Washington, NH (32%).
Method: daily TMAX, TMIN, PRCP, SNOW, SNWD from GHCN-Daily, quality-flagged values dropped; a city is included for a holiday only with โฅ20 years of data. Precipitation is the station's 24-hour total, which may not line up with evening hours. Past weather is not a forecast.