How we calculate hail report estimates

Radar data sources

We use MESH (Maximum Estimated Size of Hail), a radar product from NOAA's Multi-Radar Multi-Sensor system (MRMS). It estimates hail size on a grid of about 0.6 mile. These values aren't ground measurements or eyewitness observations. NOAA's description of MESH.

We estimate timing from NOAA's NEXRAD Level III hail detections, available through the Severe Weather Data Inventory. We use detections near the report's towns during its Eastern-time storm day. With fewer than five detections, we omit timing. We trim outlying times and round down to five-minute marks, then show the window in the lead town's time zone. These times describe hail aloft, not when it reached an address. NOAA's Severe Weather Data Inventory.

Storm dates and local time

We group radar data by calendar day in Eastern time, from midnight to midnight. In Central time, that window runs from 11 p.m. the previous evening to 11 p.m. on the named date. A late-evening storm west of the Eastern time zone can contribute to two daily maps. A link to an adjacent date means the reports share at least one town; it doesn't establish that they cover one continuous storm.

Town estimates and area maximums

Each town's figure is the maximum radar estimate within 3 miles of its center during the Eastern-time storm day. Search areas can overlap, so nearby towns can share the same maximum. A town's value doesn't describe an individual address. The date-page table and related-report links use the maximum across the whole storm area, which can exceed every town's value.

Report selection criteria

We publish a report when towns meeting the 0.5-inch radar threshold have a combined population of at least 25,000, the area has at least 15 square miles of estimates at or above 0.5 inch, and at least one town has an estimate of 0.75 inch or larger. The population total counts whole towns; it isn't a count of people beneath the swath. We group nearby towns into area reports. These criteria select reports; they don't define when roof damage starts.

Repeated radar signals

Repeated signals at the same location can be radar clutter. With data for at least six of the previous 14 days, we exclude cells with estimates of at least 0.5 inch on four or more days, plus surrounding cells. This filter can also exclude real hail. With fewer than six days available, we don't apply it.

Hail estimates and roof damage

Radar estimates show where hail may have occurred and its estimated size. They can be too high or too low, and they don't confirm hail on the ground or roof damage. Hail of 0.5 to 0.75 inch can damage roofs, especially older roofs. An estimate below 1 inch doesn't rule out damage. Inspect the roof to assess its condition.