When Does Migraine Weather Season Start? City by City

For most of the country, barometric trigger season opens in October and November. October is the first month of the year when a large block of cities crosses the line: in 36 of the 100 largest U.S. metros, at least 35% of October days carry a pressure drop big enough to count. By November that number is 58. By January it is 86.

Your own start month depends mostly on your region. Every metro we measured in the Northeast and the Midwest is in season by November. Across the South, only 13 of 42 have started by then, and most wait for December or January. Five Pacific Coast metros never cross the line in any month at all. What follows is the month by month national picture, then start and end months region by region, computed from three years of hourly pressure readings.

The national calendar, month by month

A month counts as trigger season here when at least 35% of its days record a sea-level-equivalent pressure drop of 5 hPa or more within 24 hours. That is the same definition the 2026 U.S. Migraine Risk Index uses for season length, applied one month at a time instead of totaled up.

July is empty. Not one of the 100 metros records a qualifying drop on 35% of its July days, and August manages exactly one: Anchorage, where the season runs from August through May. Summer flattens the barometer nearly everywhere, which is why the calendar reads as one long winter block with a quiet gap cut into the middle of it.

October and November turn on most of the map

Six metros are already in season as September ends: Colorado Springs, Fargo, Duluth, Boston, Providence and Anchorage. Everywhere else, the season has not started yet.

Then it arrives in two waves. Thirty metros cross the line in October after sitting below it in September. Twenty-three more cross in November.

Among the 83 metros whose season runs as one continuous block, October is the most common opening month at 25 metros, followed by November at 21, December at 16 and January at 12. Closing months bunch up harder than opening months do. April ends the season in 32 of those metros and March in 19.

Start and end months by region

Northeast. The longest seasons on average. All 10 metros are in season every month from November through March, the median season runs 8 months, and Boston opens in September without closing until June. Of the region's nine single-block metros, four run into June and four end in April.

Midwest. Twenty-three metros, not one of them without a season, median length 7 months. October opens it in 11 of the region's 20 single-block metros and November in 7. April closes it in 14 of them.

Mountain West. The widest internal spread anywhere. Colorado Springs runs 9 months, while Mesa, Phoenix and Tucson record no qualifying month at all, which pulls the regional median down to 5. Five of the seven single-block metros here open in October.

South. The biggest group at 42 metros, and the latest to start. Exactly one is in season in October and 13 by November, but 40 are by January. December opens the season in 16 of the region's 40 single-block metros, November in 12 and January in 11. March closes it in 16.

Pacific Coast. The mildest barometer here and the least consistent region. Five of its 12 metros never cross the line, so the median season is a single month, and yet Anchorage, grouped with the coast, runs 10 months and ties the longest season in the sample alongside Boston, Providence, Fargo and Duluth.

Thirteen metros, start to finish

Season windows for a spread of well known metros, at least one from every region. A window of "none" means no month reached the 35% line.

Seven metros run two separate seasons

Four of them split the same way. Minneapolis, Omaha, Rapid City and Syracuse each qualify in June, fall below the line for July, August and September, then run from October through April. On a calendar it looks like a long winter season with one stray summer month attached. Anyone tracking triggers in those four metros should know that June does not behave like the three months that follow it.

The remaining three split on different months. Flagstaff runs January through April and adds a single qualifying October. Albuquerque runs January through March, then November. Reno qualifies in February and March, and again in November and December.

Ten metros where the season never starts

Ten of the 100 never reach the 35% line in any month: Mesa, Phoenix and Tucson in Arizona, Bakersfield, Fresno, Los Angeles and San Diego in California, Tampa and Miami in Florida, and Honolulu.

A fixed threshold makes some of these closer calls than they look. Fresno peaks at 34.1% of days in February, missing the line by less than a percentage point, so its February is barely different from a qualifying month elsewhere. Honolulu peaks at 2.2% of days in January and is not a close call at all. Living in one of these ten means your barometer is quiet, not that it never moves, and whether relocating actually helps is a separate question with a more complicated answer.

What to do with a start month

A start month is only useful if pressure is genuinely one of your triggers, and that takes a record rather than a memory. The test is a few months of logged attacks set against the days the barometer actually moved, which is harder than it sounds, because the drop often arrives before the weather does. Some tracking apps pull the readings in for you: MigrAid, for example, records barometric pressure, temperature and humidity automatically with every entry on iPhone and Android, so the comparison does not rest on you remembering a forecast.

I built this app for my mom and my wife, who both get migraines. The weather side of it came from watching them try to prove a trigger they could feel and never pin down.

One caution about timing. If your attacks are becoming more frequent or harder to treat, or your pattern is changing in a way this calendar does not explain, that belongs in front of your doctor rather than in a spreadsheet.

How these numbers were made

Every figure above comes from the same source as the Risk Index: hourly surface pressure from the Open-Meteo historical archive, covering April 1, 2023 through March 31, 2026, for the 100 largest U.S. metros. Readings are scaled to a sea-level equivalent, so a swing in Denver, where mean pressure measures 838.4 hPa, is compared fairly against the same relative swing at the coast.

A day counts when the worst 24 hour change ending in it is a drop of 5 hPa or more. A month counts as season when 35% of its days qualify. The report's methodology section carries the full detail.

The 5 hPa figure is not ours. In a year long diary study of 28 migraine patients published in the journal Internal Medicine in 2011, Kimoto and colleagues found headache frequency rose when pressure fell by more than 5 hPa from the day of the headache to the day after, and weather change was associated with attacks in 18 of the 28.

Two limits are worth stating plainly. A 2024 review in Current Pain and Headache Reports put weather's effect on migraine attacks at around 20%, called recent study results inconsistent, and reported no causative relationship between weather and migraine patterns.

Beyond that, this is a climate calendar and not a forecast. It describes what the barometer has usually done in your metro during a given month across three years, not what it will do next week, and not whether you will have an attack. Weather is one trigger among many, and for most people it is not the largest.

For the coming few days rather than the coming months, the free risk forecast pages cover 170 U.S. cities and show pressure measured at each city's own elevation, which is the number your body experiences rather than the sea-level value most weather apps display.