Hail claim denied? How to prove the date of loss

Most hail claims are not denied because the roof looks fine. They are denied, delayed or cut down because the insurer is not convinced the damage happened on the date you reported, at your address, while the policy was in force. That is a question about time and place, and it can be answered with records. This guide explains why the date of loss becomes the fight, which public weather records help, where each one falls short, and how to put together a file an adjuster or an attorney can check.

Why does the date of loss matter so much?

An insurance policy covers a specific period. If hail damaged your roof two years ago and you bought the policy last spring, the damage is not covered, and insurers know that old hail damage often gets reported after a new storm. That is why insurers check reported dates against weather data. Verisk, for example, sells Benchmark date-of-loss reports to the claims industry; its published lightning sample lists every recorded strike within 5 miles of a property in the 36 hours either side of the reported date.

So if your claim was denied with language like no storm activity at the risk location on the reported date, or damage is consistent with wear and prior events, the insurer is usually not saying your roof is undamaged. It is saying your date is not supported.

• The wrong dayPeople often notice hail damage days or weeks later and report the date they found it, not the date it happened. If a stronger storm hit three days earlier, that is the date the records will support.
• The wrong stormIn an active season several storms cross the same area. Reporting the one you remember, rather than the one that dropped the largest hail on your street, weakens the claim.
• Prior damageIf the roof or car already had hail marks, the adjuster has to separate old from new. A clear date for the new event helps draw that line.
• Late noticeYour policy may set a deadline for submitting hail and wind claims, and some states limit when an insurer can deny for late notice. In Colorado, for example, United Policyholders explains that an insurer may not deny a claim for untimely notice unless it can show it was prejudiced by the delay.

The free public records, and what each one proves

Start with the free sources. They cost nothing, and an adjuster will check them anyway.

NOAA Storm Prediction Center storm reports

The SPC publishes a daily list of local storm reports sent by National Weather Service offices, with location, time and hail size. Each day runs from 12:00 UTC to 11:59 UTC the next day, which is early morning in the continental US, so an evening storm can appear on the next date's page. The SPC itself says the reports are preliminary and likely do not contain every severe weather report for an event. A storm that hit a rural road with nobody to report it may simply not be there.

NCEI Storm Events Database

The Storm Events Database is the official, final record of severe weather reports. According to the SPC FAQ, it takes a few months to compile, so for a recent storm it may not have your event yet. For an older claim it is the strongest public source on whether a reported event happened in your county.

Radar: MRMS hail size estimates

NOAA's Multi-Radar Multi-Sensor system estimates hail size from radar, the MESH product. It covers the whole country on a fine grid, which makes it useful far from any spotter. But MESH is an estimate, and researchers at NSSL describe it as meant to represent roughly the 75th percentile of observed hail diameters, not the largest stone that fell. Read it as storm intensity over your area, not as a measurement of what hit your roof.

A real example: Kankakee, Illinois, March 10, 2026

Our published US sample report covers an address on South Osborn Avenue in Kankakee, Illinois, with a 1 km radius, for March 10, 2026. It shows what a date-of-loss file looks like when the sources agree, and why they never read exactly the same.

• Storm reportsThe SPC list for that period carried 35 hail reports. The nearest, 0.2 km from the address at 23:28 UTC, describes a 6.14 inch hailstone found on Osborn Avenue. The report is flagged as preliminary.
• RadarMRMS MESH peaked at 54.6 mm inside the analysis circle and 59 mm within 5 km. That is a giant-hail class reading, and it is still far below the 6 inch stone, which is exactly what you would expect from a 75th-percentile estimate.
• SatelliteGOES-19 measured cloud tops as cold as −75.4 °C over the area, and the multispectral hail reading reached a 77.5 % probability on the point.
• LightningThe GOES-19 lightning mapper recorded 2 flashes inside the 1 km circle during the window.

Four independent records, one place, one evening. That is the kind of file that ends a date-of-loss argument.

What a satellite report adds

Spotter reports depend on someone being there. Radar depends on beam height and distance from the nearest site. A geostationary satellite watches the whole storm every few minutes from above. GOES-19 measures cloud-top temperature, ice-particle signatures and, with its Geostationary Lightning Mapper, lightning activity, and it does so whether or not anybody filed a report.

A Skybridge report pulls those measurements for your address and date and puts them next to the NOAA radar estimate and the SPC reports within 50 km. You get a verdict with a probability, maps, the hour the storm peaked over you in local time, the methodology and every source cited. Each report carries a unique code so the adjuster can open the original version on our site and confirm nothing was edited.

• It answers the time questionThe satellite sequence shows when the storm was over your address, to the 10-minute step.
• It answers the place questionThe analysis is centred on your address, with the radius you choose.
• It answers the which-storm questionIf you are unsure of the date, you can add up to two more days and see which one the evidence supports.
• It does not assess the roofNo weather record can tell whether a dent came from this storm. That is the inspector's job. The report tells the inspector what the sky did.

Building your date-of-loss file

• 1. Pin down the dateCheck SPC reports and, for older claims, the Storm Events Database for your county. If two storms are candidates, get the evidence for both.
• 2. Add your own recordsDated photos and video of the hail and the damage, receipts for emergency repairs, neighbours' claims and contractor reports.
• 3. Add an address-level reportOne document that puts satellite, radar and storm reports for your exact location on one page, with a verification code.
• 4. Respond in writingAsk for the denial reasons in writing, then answer each point with the evidence. Keep copies of everything you send.
• 5. Know your next stepsIf the insurer will not move, an appraisal clause, a public adjuster, your state's insurance department or an attorney may be next. Policy terms and state rules differ, so check yours.

A US report starts at USD 249. Coverage starts in 2018, so it works for older claims still in dispute. If you also need to document lightning damage, see our guide Lightning strike report for an insurance claim, and for a comparison of every type of storm evidence, read Satellite, radar or ground network: which storm evidence holds up?