The American Hail Belt 2016–2025

What is in here

This is the whole NOAA Storm Events hail record for the United States over ten years, aggregated to the county and published as a table you can read, sort, and quote. 3,026 counties across 50 states. 87,206 hail events and 172,394 severe-wind events. The largest stone on the record measured 5 inches.

We publish it because we already had to build it. Curbsight joins this record to property data to work out which roofs sat under which storms, and the county aggregate is a by-product of that. There is no reason to sit on it. The data is public, the aggregation is ours, and a contractor deciding whether to open a branch in Amarillo should not have to write a parser to find out what the weather has been doing there.

Take the numbers. Cite the page. If you are a journalist working a storm and you need the ten-year context for a county, it is in the table.

The national picture

Ten years, 87,206 recorded hail events. 19,666 of them reached 1.75 inches — golf ball — which is the rough line where hail stops being noise on a gutter and starts being a claim. That is 22.6% of everything recorded, or one event in 4.4.

The 1.75-inch line is not our invention and it is not a rule. It is a diameter NOAA measures, and it is the size the trade and most carriers treat as the point where a composition roof is plausibly damaged. Plenty of roofs get destroyed by smaller stones falling for longer, and plenty of roofs shrug off a single golf ball. It is a threshold, not a verdict.

• 3,026 counties have at least one recorded hail event in the last decade — that is most of the country, not a belt.
• Only 37 states record enough hail for a severity rate to mean anything. The rest are in the totals but not in the ranking.
• Severe wind outnumbers hail almost two to one nationally, and it is the damage category contractors most often leave on the table.

Frequency is not severity

Here is the part that changes a decision.

The instinct when picking a market is to ask where it hails most. That question has an obvious answer — Texas, by a distance, with 12,900 recorded events. But the useful question is what happens when it does hail, and those two answers do not line up the way you would expect.

Texas converts 31.4% of its hail to claim size. Nebraska records 41% of Texas's hail volume and converts 29.4% of it — nearly the same rate, on a fraction of the events. Meanwhile there are states well up the volume ranking that convert under 20%. Plot the two against each other and there is no clean diagonal. Volume tells you how often your phone rings. The severity share tells you how many of those calls are worth driving to.

Neither number is the whole story and we are not going to pretend it is. A market is territory, competition, carrier behaviour, and whether you can staff it. But if you are going to reduce it to weather, reduce it to the right weather.

By state

Ranked by claim-sized events, because that is the column that correlates with work. The rate column is blank for states below 250 recorded events — not because they get no hail, but because a percentage built on a small count says more about the denominator than the weather.

By county

The twenty-five counties with the most claim-sized hail in the country over the decade. This is the list worth pinning to a wall if you are deciding where to put a crew.

Each one links to its full storm history — every documented event, the notable ones by date and size, and the severe-wind record alongside it.

What this data cannot tell you

NOAA counts reports, not storms. A hail event enters the database because a person or a spotter network reported it, which means the record follows population. A county with three thousand people and a bad storm season can look quieter than a suburb that got pea-sized hail and has an active weather club. That bias lands almost entirely on the frequency numbers.

It is the reason we lead with the severity share rather than the raw count. The share of reports that were large is far less sensitive to how many people were around to make a report in the first place. It is not immune — nothing here is — but it is the sturdier of the two.

The other thing the record cannot tell you is what happened to any particular roof. A county with 400 documented hail events does not mean 400 damaged neighbourhoods, and a roof under a golf-ball storm in 2019 may have been replaced in 2020. Storm history is where the conversation starts. It is not the conversation.

What to do with this on Monday

Three things, in order, and the first two cost nothing.

• Look up the counties you already work. If the ten-year severity share is below the national 22.6%, your storm pipeline is thinner than your instinct says and your business needs a second leg — retail, service, or a recurring line.
• Look up the county you have been thinking about expanding into, and compare its severity share to the one you are in now, not to its raw event count. The count is the number that will mislead you.
• Pull your own last five years of jobs and mark which ones came from a storm. Compare that ratio to what the weather record says the county gave you. If the weather was there and the work was not, the gap is a coverage problem, not a demand problem — and that one is fixable.

Where this comes from

The source is the NOAA Storm Events Database, the federal record of severe weather reports. The aggregation, the county roll-up, and the severity share are ours. The dataset behind this page is the same one that powers our free hail-history lookup, which serves all 3,026 counties rather than the twenty-five printed here.

If you want to see what this looks like joined to actual properties — which roofs sat under which storm, and when they were last replaced — that is the product, and the honest way to evaluate it is to ask us to run your own county and check the answer against what you already know about it. Bring a market you know cold. That is the only test that means anything.

Questions