Storm Stream.

science

Why Hail Size Decides Roof Damage

Lab work on asphalt shingles says stone diameter matters, but roof age and the kind of damage matter more than the 1 inch rule suggests

The 1 inch line was drawn for warnings, not for roofs. The Storm Prediction Center defines severe hail as hail 1 inch in diameter or larger, and significant severe hail as 2 inches or larger. The National Weather Service uses the same figure in its severe thunderstorm criteria, where 1 inch is quarter size. Those numbers decide when a warning goes out and when a report lands in a database. They were never written to tell anyone whether a roof needs replacing.

The field borrowed them anyway. One inch became the line between a storm worth chasing and a storm worth ignoring. Two inches became shorthand for a total loss. Two pieces of lab work suggest both halves of that habit are shakier than they look, and they fail in opposite directions.

The size chart people actually carry

Hail size gets reported by comparison to a familiar object, which is why the reference chart is the real working document. The NWS Burlington SKYWARN reference gives the standard pairings:

The same page notes that most local hail reports are 3/4 inch or smaller. Keep that in mind, because the stones under the severe line are the ones that fall most often.

What the lab says about new shingles

IBHS tested commercially available asphalt shingles with 1.5 inch and 2 inch laboratory hailstones, bought new through the typical consumer supply chain. Every product tested with 1.5 inch laboratory hail rated Good or Excellent. In their new state, both impact-rated and non impact-rated asphalt shingles should offer good protection against 1.5 inch hail or smaller. Impact-rated means Class 3 or 4 under the UL 2218 or FM 4473 testing programs. Everything else is not impact-rated.

Impact-rated products generally offered about 16 percent better total performance scores. At 2 inches the picture splits. Where the impact-rated shingle beat the product that was not impact-rated against 2 inch impacts, performance scores for impact-rated products rated Good and Excellent averaged 14 percent higher, and for some individual manufacturers and products the improvement ran as high as 34 percent. Every impact-rated shingle that rated Good or Excellent against 2 inch impacts was polymer-modified asphalt.

IBHS states that the lab stones' impact energy and material properties are directly related to typical natural hailstones for those sizes in non wind-driven conditions. It also says plainly that it did not test all available asphalt shingles, and that it is a nonprofit research organization funded by the property insurance industry. Both of those belong in any argument built on this data.

What the lab says about small stones arriving in volume

A study published in Frontiers in Materials on September 29, 2025 went after the stones nobody tests. The authors fired ice spheres of 17.8 mm, about 0.7 inch, and 25.4 mm, about 1 inch at shingle specimens at a concentration of 44 impacts per square foot. They picked that concentration from available field data as a realistic high-concentration scenario, and they note that no standard threshold exists for it. Each test series used 300 of the smaller spheres and 200 of the larger, so each specimen absorbed 1,000 impacts across two series.

The headline result: the cumulative granule loss from the sub-severe impacts in test series 1 exceeded the total granule loss of a single 50.8 mm impact, which is 2 inches. Earlier work had suggested that hailstones smaller than 25.4 mm rarely cause functional damage.

Per impact, the small stones are still small. Table 6 of that paper reports 9.5 square mm of granule loss per sub-severe hit on the control specimens and 10.2 square mm on the experimental specimens, against 20.0 square mm for a single 2 inch impact on new product. That is the whole mechanism in one comparison: roughly half the loss per hit, delivered hundreds of times instead of once.

Age moves the number further than size does

The same study included a naturally weathered group, and that is where the figures separate. On naturally weathered specimens, 50.8 mm impacts averaged 192.2 square mm of granule loss per impact, against 57.2 square mm for the control and 20.0 square mm for the new-product baseline. Dividing 192.2 by 20.0 gives a factor of about 9.6 for the same stone size on an aged surface rather than a new one.

The study reports an average performance decline of 47 percent, with a range of 39 percent to 67 percent. It also reports a Mann-Whitney U test with U = 2.00 and p = 0.0087, comparing median granule loss of 16,503 square mm in the experimental group against 9,287.5 square mm in the control.

Granule loss is not the damage that leaks

Functional damage, as that study defines it, means punctures or fractures in the shingle that compromise its water-shedding capability, or damage that otherwise reduces service life. A bare spot is not automatically either of those.

IBHS found the same divide from the other direction. New products all received Good and Excellent granule loss scores, while dents and ridges were consistently the poorest scores. And IBHS identifies dents, ridges and tears as the damage types most directly correlated to water penetration. So a roof can shed granules visibly and still shed water, and a roof can look lightly scuffed while carrying the dents that matter.

Where both studies stop

Read the limits before quoting either one. The 2025 study tested only six products, two specimens each, and the authors call the sample small and say more testing is needed. Both studies used manufactured ice spheres rather than natural stones, which differ in shape, density and fall angle.

The 2025 paper also disagrees with itself. Its text describes the sub-severe stones as causing nearly 30 percent of the granule loss of a 2 inch stone, which does not match its own Table 6 figures, where 9.5 and 10.2 square mm against 20.0 square mm works out to roughly 48 to 51 percent. The Table 6 numbers are the ones to use. The discrepancy is worth knowing about if someone across the table has read the summary and not the tables.

What a size number can and cannot settle

A diameter tells you how much energy was available. It does not tell you what the surface it hit was able to absorb.

Two things follow. The same reported stone size means different things on a two-year-old roof and a ten-year-old roof, and the weathering figures above are large enough that age deserves to come up before size does. And a storm that produced only sub-severe hail is not automatically a storm that produced no damage, particularly if it produced a lot of it for a long time.

None of that makes a size number useless. It makes it one input among several. What decides whether water gets in is whether the mat is punctured, dented, ridged or torn, and the only way to establish that is to go look at it.

Sources

More on this