Storm Stream.

science

When Hail Season Peaks, Region by Region

Hail season is a wave that moves north through the year, and the report records show when each region peaks

Ask when hail season is and you get an answer that is true for the country and wrong for your state. Most states see their largest fraction of annual severe hail reports in May, June or July, according to the USDA Midwest Climate Hub's 2024 factsheet on hail in the Midwest. Break that apart and the single season turns into a wave that moves.

The peak is not a date, it is a wave moving north

Southern states such as Missouri and Illinois see their highest monthly hail fractions during April and May. Northern states such as Wisconsin and Minnesota, along with the Northern Plains, run later, with higher monthly hail fractions in June, July and August.

The measure the factsheet uses is monthly hail fraction: the percentage of a state's annual hail reports that fall in a given month, drawn from the NCEI Storm Events Database. It is a share, not a count. It answers when, and says nothing about how much.

That distinction matters for anyone who staffs by season. A state whose peak share lands in April is not necessarily a state with more hail. It is a state whose hail arrives earlier, and the useful consequence is that a crew can work the Gulf and lower Mississippi valley early and still be in Minnesota for its own peak.

Two local records, with their dates attached

Local office climatologies go deeper than the national picture, and they are old. Say so when you use them.

NWS Lake Charles, Louisiana compiled 404 hail reports in its County Warning Area over the period 1955 to 1995. In that record, 67 percent of hail reports occur during the March through May period, and May is the peak month with 105 reports, which is 26 percent of all reports. The same study notes a secondary peak in the autumn months.

For Fort Smith Municipal Airport, NWS Tulsa published a record covering 1972 through December 1997: 50 total hail occurrences, about 2.0 hail events per year. May led with 14 occurrences, a 56 percent monthly chance, followed by April with 11 occurrences at 44 percent and March with 10 at 40 percent. A secondary peak shows up in November with 4 occurrences, a 16 percent monthly chance.

Two things about those numbers. They end in the 1990s, so they describe the second half of the twentieth century and not the present. And the Fort Smith record is 50 events at a single airport, which is a thin enough sample that a month or two of difference should not be read as a trend.

Where the volume actually lands

Share tells you when. Counts tell you where. The Insurance Information Institute reports 5,432 hail events in the United States in 2025, up from 5,373 in 2024, citing the Storm Prediction Center's annual severe weather report summary. The top states by hail events in 2025:

Texas alone carries more events than the next two states combined, which is arithmetic on those figures: 902 against 375 plus 369. The same page reports that State Farm hail claim payouts in 2025 were over 5,600,000,000 dollars, up more than 600,000,000 dollars from 2024, and that in 2024 the company paid 3,800,000,000 dollars for home repairs and over 1,200,000,000 dollars for auto repairs.

Set the share data and the count data side by side and the planning question changes shape. Tennessee and Illinois both appear in the top ten by count, and they sit on different sides of the seasonal wave.

Why a report count is not a climatology

Every number above the local-office level comes from reports, and reports have known biases that the agencies publishing them describe openly.

The Storm Prediction Center's severe weather database, compiled from NWS Storm Data, carries hail and damaging wind reports back to 1955. Reports cluster around population centers. The data are used for NWS warning verification and may not capture all storm events. The current year's reports are preliminary. Monetary loss figures in the database are described as highly suspect.

There is also a threshold break in the middle of the record. The NWS raised the severe hail threshold from 0.75 inch to 1.00 inch in 2010, and for legacy purposes 0.75 inch reports remain in the annual hail files, so anyone pulling a long time series may need to filter. A year-over-year comparison that straddles 2010 without that filter is comparing two different definitions.

The Midwest factsheet says the same thing in its own words: assessing hail climatology is difficult largely because the reporting of hail has been inconsistent.

The trend question, answered carefully

Comparing 1991 to 2020 against 1961 to 1990, averaged across all states, the factsheet finds that monthly hail fraction slightly increased in April and May, decreased in June and July, and increased from August through November. It then qualifies that directly: the changes are minimal, about 2 percent or less, and are not statistically significant for most states.

Two other measures in the same document point the same direction. Crop insurance hail indemnity payments over 2014 to 2023 are highest in June, July and August for most states, and comparing back to 1989 showed no clear shift. From 1979 to 2017 there was an increase in environments favorable to large hail by 2 to 4 days in the Midwest, which is a change in ingredients rather than in observed hail. For scale, hail crop losses amount to roughly 5 percent of crop value annually.

The honest summary is that the seasonal shape has not moved much, and the data are not clean enough to say it has.

Autumn, and the second peak

Both local records show a secondary autumn peak: Lake Charles in the autumn months generally, Fort Smith specifically in November. The newer-period comparison in the Midwest factsheet also shows monthly hail fraction increasing from August through November, small and not statistically significant though it is.

None of these three sources explains the mechanism behind a second peak, and guessing at one is not worth the risk of being wrong in print. What they do establish is that the autumn window is real in the records, modest, and most visible in the south-central states.

The practical weight of that is scheduling. A 16 percent chance of hail in November, which is what the Fort Smith record from 1972 through 1997 works out to, is not a second season. It is enough to argue against treating late autumn as dead time in Arkansas or Louisiana, and enough to make a November storm there unsurprising rather than anomalous.

What to take from this

If you work one state, the national May-through-July answer is not your answer. Pull the share by month for your own states, then pull the counts, and treat them as answers to two different questions.

If you work several states, the wave is the whole opportunity: earlier in the south-central states, later in the upper Midwest and Northern Plains. If you are comparing years, check whether your series crosses 2010 and whether 0.75 inch reports are still in it. And when you cite a local office climatology, cite its period of record in the same sentence. A record that ends in 1995 is evidence about 1955 to 1995.

Sources

More on this