Why there is no single threshold
Every homeowner wants one number: the diameter above which a roof is damaged and below which it is fine. That number does not exist, and anyone who quotes it without qualification is selling something.
The reason is that hail damage is an energy problem with too many inputs. Kinetic energy rises with the square of impact speed and with the cube of stone diameter, so a small change in size is a large change in energy. But the energy delivered is only half the equation. The other half is how much energy the roof can absorb before something breaks, and that varies by material, by age, by temperature, by how stiff the deck is beneath the shingle, and by the angle at which the stone arrives.
What we can do is describe bands. Within a band, the range of likely outcomes narrows enough to be useful — enough to tell you whether to ignore it, look at it, or call someone.
The variables that decide it
- Material. Aluminum gutter dents at sizes that do nothing to a shingle. Clay tile chips at sizes that only scar asphalt. A single-ply membrane over soft insulation fractures at sizes that a membrane over a dense cover board shrugs off. See roofing materials and how they handle hail.
- Age. The dominant variable on asphalt. Shingles oxidize and stiffen year on year, so the same stone that leaves a granule scar on a four-year-old roof can crack the mat of a seventeen-year-old one. Work out where you sit with the roof age calculator.
- Temperature. Cold asphalt is stiffer and less able to deform without fracturing, so a spring storm at dawn is harder on a roof than the same storm on a hot afternoon.
- Wind speed and impact angle. Wind adds horizontal velocity, raising total impact energy and turning a vertical strike into an angled one that concentrates load on a smaller contact area.
- Stone density and shape. Hailstones are not uniform ice spheres. Some are dense and clear, some are low-density and full of trapped air, and many are lobed or spiked. A dense stone with a protruding lobe delivers its energy through a small area and does disproportionate damage for its nominal diameter.
- Deck stiffness. A strike over a rafter or over stiff sheathing has nowhere to go and the shingle absorbs it all. Over a flexible or spongy span, the deck deflects and takes some of the load.
The practical size bands
Below 1 inch. Rarely produces functional damage to a sound asphalt roof. It does knock granules loose, mark soft metals lightly, damage screens and can crack acrylic skylight domes. This size band is also where the cumulative story lives: a metro that gets pea-to-nickel hail several times a year will wear its roofs out early even though no single storm was claimable.
1 to 1.5 inches. The contested band, and the one that produces most claim disputes. On a newer, well-installed architectural roof this often leaves granule scars without fracturing the mat. On an old, brittle, or 3-tab roof, bruising becomes common at the top of this range. Soft metals are clearly dented, which is why the corroborating evidence in hail damage vs normal wear matters so much here — the metal proves the storm, the thumb test proves the roof.
1.5 to 1.75 inches. Bruising on aging asphalt is likely; damage on newer asphalt becomes plausible. Wood shakes split. Older or thinner tile starts cracking.
Above 1.75 inches. Functional damage to asphalt is common rather than exceptional, across roof ages. Vents, flashings, ridge caps, skylights and siding start taking damage in their own right.
Above 2.5 inches. Damage is usually obvious from the ground. Expect punctured or torn shingles, cracked and displaced tile, damaged decking in the worst cases, destroyed vents and turbines, dented garage doors, and impact damage to siding and window frames on the windward elevations.
They are a synthesis of how the materials behave, not the output of a dataset that maps stone size to outcome for your specific roof. Nothing here overrides what an inspection finds. A 1.25-inch storm can total a twenty-year-old roof, and a 2-inch storm can leave a two-year-old impact-rated roof with nothing but scuffs.
Hail size, object name and what it affects
Storm reports and radar products both express hail size in inches by diameter, and the National Weather Service uses a standard set of everyday objects so that public reports are comparable. One inch — quarter size — is the severe criterion for a warning, which is a threshold for public safety messaging rather than a statement about roofs.
| Inches | cm | Object | Typically affects |
|---|---|---|---|
| 0.25 | 0.6 | Pea | Nothing structural; leaves, garden plants |
| 0.50 | 1.3 | Marble or mothball | Loose granules move; no roof damage expected |
| 0.75 | 1.9 | Penny | Granule displacement, light screen and fin marks |
| 0.88 | 2.2 | Nickel | Soft-metal dimples begin; skylight domes at risk |
| 1.00 | 2.5 | Quarter — severe criterion | Clear soft-metal denting; granule scars on asphalt |
| 1.25 | 3.2 | Half dollar | Bruising possible on old or 3-tab asphalt |
| 1.50 | 3.8 | Ping-pong ball or walnut | Bruising likely on aging asphalt; wood shakes split |
| 1.75 | 4.4 | Golf ball | Functional asphalt damage common; vents and flashing hit |
| 2.00 | 5.1 | Hen egg | Widespread mat fracture; tile cracks; gutters deformed |
| 2.50 | 6.4 | Tennis ball | Obvious damage; siding, windows, garage doors, skylights |
| 2.75 | 7.0 | Baseball | Punctured shingles, broken tile, damaged decking possible |
| 3.00 | 7.6 | Teacup | Severe roof and envelope damage; vehicles totaled |
| 4.00 | 10.2 | Grapefruit | Structural roof damage; penetrations through covering |
| 4.50 | 11.4 | Softball | Catastrophic; decking, sheathing and interior exposure |
Radar-estimated size vs what landed
The hail sizes on our address lookup come from NEXRAD Level-III hail detection output archived by NOAA. The radar does not see hailstones. It sees reflectivity and its vertical structure, and an algorithm infers from the height and intensity of the reflectivity core how much hail the storm is likely producing and how large the biggest stone probably is. The output includes a probability of hail, a probability of severe hail, and a maximum estimated size — the field usually labeled MAXSIZE.
Three things follow from how that number is produced, and they are the difference between using it well and misusing it.
- It is an estimate of the largest stone aloft in the cell, not the typical stone. Most of what falls in a hail swath is smaller than the maximum. A cell credited with 2 inches may have dropped mostly half-inch stones over most of its footprint.
- It describes the cell, not your address. Radar sample volumes are large and grow with distance from the site. Ground truth varies enormously across a few hundred metres — one street gets golf balls, the next gets pea hail, and the swath edge can run down the middle of a block.
- It says nothing about wind, angle or duration. Two storms with the same estimated size can produce completely different damage depending on how hard the wind was driving and how long the core sat over the house.
A large estimated size near your address means the storm was worth inspecting. It is not evidence that your roof was damaged, and a small estimate is not evidence that it was not. Full detail on how these signatures are derived and their known limitations is on the methodology page.
Ground reports are the useful complement. Trained spotter and public reports of actual stone size, collected by the National Weather Service and compiled by the Storm Prediction Center, tell you what people picked up off the lawn. They are sparse and biased toward populated areas and daylight hours, but where one exists near you it is worth more than any algorithm output.
Wind, angle, density and shape
A stone falling in still air reaches a terminal velocity set by its size and density. Add a 60 mph horizontal wind and the resultant impact velocity rises, and — more importantly — the angle changes. An angled strike concentrates load on a smaller effective contact patch and drives more damage into the windward slope while sheltering the leeward one. That is why damage patterns are directional, and why a roof marked evenly on all four slopes usually is not telling a hail story at all.
Shape matters as much as size. A spiked or lobed stone of nominal 1.5-inch diameter can deliver its energy through a contact patch the size of a fingertip. Density matters too: stones grown in strong updrafts with lots of accreted clear ice are heavier for their diameter than soft, milky, air-filled ones. None of this is captured by a single diameter number, which is exactly why size alone never settles the question.
How to use a size number sensibly
Treat hail size as a filter, not a verdict. Look up the storms over your address on the hail history lookup, note the dates and the estimated sizes, and use them to decide where to spend attention. If a 2-inch signature passed over you eighteen months ago and your roof is fifteen years old, that is worth an inspection and a set of dated photographs. If nothing over an inch has been near you in five years and the roof looks tired, the roof is tired.
Then work the evidence in the right order: photograph the soft metals with the photo checklist, apply the diagnostic tests in hail damage vs normal wear, and only then get into the claims process on the insurance claims pages. If you are replacing anyway and you live somewhere that gets hit repeatedly, impact-resistant shingles is the next thing to read, because moving the whole damage curve up a band is the only durable fix.
Common questions
What size hail is officially severe?
The National Weather Service treats hail one inch in diameter -- quarter size -- and larger as severe. That is a warning threshold for the public, not a roof damage threshold.
Can hail under one inch damage my roof?
It can strip granules, dent soft metals, crack skylight domes and mark vents, and on a brittle old roof it can do more. Functional damage to sound asphalt below one inch is uncommon.
Why do contractors and insurers disagree in the one to one and a half inch range?
Because that is exactly the band where roof age, material and temperature decide the outcome. The same storm produces a claimable loss on one house and cosmetic scarring next door, and both opinions can be defensible.
Does radar tell me how big the hail was at my house?
No. Radar estimates the largest stone the algorithm believes was aloft somewhere in the storm cell. Actual stone size at ground level varies enormously over a few hundred metres, so treat it as a prompt to inspect, not as evidence.
Does hail size alone determine damage?
No. Wind speed, impact angle, stone density and shape, roof material, roof age, temperature and deck stiffness all change the result for a given diameter.
What size hail damages a metal roof?
Metal dents at fairly small sizes and keeps shedding water long past the point where asphalt has failed. Functional damage to steel panels generally needs large stones, though whether dents are covered depends on your policy.
Sources
Every factual claim on this page that comes from outside our own data is sourced below. Where a number is a model or a range rather than a measurement, it is labelled as one in the text.
- NOAA NCEI Severe Weather Data Inventory -- NEXRAD Level-III hail signatures including maximum estimated size
- NOAA National Weather Service -- severe hail criteria and the standard hail size to object comparisons
- NOAA Storm Prediction Center -- severe weather reports including ground-truth hail reports
- Insurance Institute for Business and Home Safety -- field research on hailstone properties and roof damage
- Insurance Information Institute -- background on hail losses and claims