Each report is a PDF that states what official weather data shows for a specific property on a specific date. We read NOAA's radar and gridded weather products at the address, cross-check them against ground observations logged by trained spotters and the public, and tell you how much confidence the combination deserves — including when it deserves very little. Carriers, adjusters and policyholders work from the same facts. Reports are typically delivered in about half a minute.
What these reports are not: a measurement at your roof. Radar estimates hail size from above; it does not touch a hailstone. We say so on every report, and we tell you when the radar's view of your address was poor.
We read NOAA MRMS MESH radar — the National Weather Service's gridded hail-size estimate — across the date of loss, and report the value at the property's own grid cell and the peak within a half mile, a mile, three and five miles, each labelled separately. We check NOAA's Radar Quality Index to confirm the radar could actually see the address, and cross-check against official hail reports nearby.
Flagged when radar indicates hail of 0.75″ or larger near the property. MESH is calibrated as a 75th-percentile upper estimate, so we lead with how likely damaging hail was, not a single number presented as fact.
We take the peak measured gust from the nearest official ASOS/AWOS stations, report how far away each station is, and keep it separate from reported gusts logged nearby, which may be estimates. Every number on the report is attributed to the instrument or observer that produced it.
Flagged when official data shows a gust of 58 mph or more — the NWS severe-thunderstorm criterion — measured at a station or logged in an official NWS storm report. A downburst is often only a mile or two wide, so a distant station recording nothing is weak evidence that nothing happened, and we grade it accordingly.
We read NOAA SNODAS modelled snow depth on the ground and snow-water-equivalent at the property, convert the water-equivalent into estimated roof load (lb/ft²), and check both against the nearest official NWS climate site's measured depth and new snowfall.
Flagged when modelled depth reaches 6″ or more. Depth on the ground includes older snowpack — it is the right measure for roof load, and it is not a statement that snow fell on the date of loss.
A second page covering official hail reports within five and ten miles over the previous 24 months, whether an NWS severe thunderstorm warning covered the property that day, and the peak measured wind gust the same day. Prior hail is history, not evidence of damage to a particular roof, and the report says so.
High — the finding is corroborated. Either the data met the threshold and independent observations agree, or the reading is clear and the instruments demonstrably had a good view of the address.
Moderate — the data supports the finding but something is unconfirmed: no nearby ground report, only fair radar or station coverage, or a model estimate with no local measurement to check it against.
Low — the signals conflict, or coverage was too poor to rely on. Ground reports exist but the estimate at the property doesn't support them, or the nearest instrument was too far away to be meaningful.
Where coverage is missing entirely, we state that and decline to give a number. An absence of data is never reported as an absence of weather.