# How radar hail estimates work, and how far to trust them

> Radar hail sizes in RoofQuery storm reports come from NOAA's Multi-Radar Multi-Sensor Maximum Estimated Size of Hail (MESH), which estimates the largest hailstone in a storm from radar reflectivity above the freezing level, on a grid about 1 km across. It is an estimate of hail aloft, not a measurement on the ground: research has found it overstates small hail and understates large hail. Storm reports therefore pair radar with ground reports, warnings and lightning, and say how strongly the sources agree.

Category: How it works. Last updated 2026-10-17.
Canonical: https://roofquery.com/guides/how-radar-hail-estimates-work

The radar hail sizes in a storm report are estimates, from a NOAA product called MESH. They are the best continuous map of where hail fell, and they are wrong in predictable ways, which is why a storm report never relies on radar alone.

## What is MESH?

Maximum Estimated Size of Hail, part of NOAA's Multi-Radar Multi-Sensor system (MRMS). It combines the national network of weather radars and estimates the largest hailstone in each part of a storm from how strongly the storm reflects radar above the freezing level, where hail grows. The [MRMS MESH product](https://journals.ametsoc.org/view/journals/wefo/36/2/WAF-D-20-0158.1.xml) is what the National Weather Service and the Storm Prediction Center use to diagnose hail size in real time. It is computed every few minutes on a grid about 1 km across; storm reports use the largest value each grid cell reached in the 24-hour storm day.

## How accurate is radar hail size?

Good at where, weaker at exactly how big. MESH measures hail in the cloud, not on the ground: stones melt on the way down, wind carries them sideways, and the radar sees a volume, not a roof. A [comparison with nearly 6,000 hail reports](https://www.researchgate.net/publication/331426504_Severe_Hail_Fall_and_Hail_Storm_Detection_using_Remote_Sensing_Observations) found that MESH tends to overstate small hail and understate large hail. Treat a radar size as the likely range at that spot, not a measured stone.

## Why does a storm report need ground reports too?

Because they measure what actually landed. Storm reports draw on National Weather Service storm reports from trained spotters and the public, and on [CoCoRaHS](https://www.cocorahs.org/Content.aspx?page=hailcoco) volunteers whose foam hail pads record the dents real stones made. Reports are precise but sparse: they cluster near towns and roads, so no report near a house does not mean no hail. Radar fills the gaps between reports; reports anchor the radar.

## How do storm reports decide confidence?

By how many independent sources agree, and how close to the property they are.

| Confidence | Typical evidence for hail |
| --- | --- |
| High | Radar hail over the property and a ground report within 3 miles |
| Medium | One strong source, such as an NWS warning over the property or a nearby report |
| Low | Radar only, or evidence well away from the property |

Each storm day also carries its evidence in plain English, for example "radar estimated 1-inch hail over the parcel; an observer measured 1.25 inches 0.2 miles away; a warning over the property called for 1-inch hail." The [storm report data](https://roofquery.com/docs/storm-report-data) reference lists every field.

## What is done to clean up the radar data?

Two things. Radar areas with no lightning nearby are removed, which drops returns that look like hail but come from something else, and values are capped at 5 inches, in line with the Storm Prediction Center's practice. Radar hail is then contoured into size bands, the coloured swath on the map, so you can see whether the property sat in the middle of the hail or on its edge.

## How should I use radar hail in a claim or a sale?

As evidence of where and roughly how big, alongside the rest. A property inside a 1.5-inch band with a 1.75-inch report a mile away is a strong case; a property on the edge of a 1-inch band with no reports is a weak one, and the confidence rating says so. Pair it with the [hail size thresholds](https://roofquery.com/guides/hail-size-chart-and-roof-damage) for roofing, and with an inspection for anything that has to prove damage.

## Frequently asked questions

### How far back does radar hail data go?

RoofQuery storm reports use NOAA MRMS radar hail from October 2020, which is why their period is at most five years.

### Is radar hail the same as hail on the ground?

No. MESH estimates the largest hail inside the storm above the freezing level. Stones melt and drift as they fall, so ground sizes and positions can differ.

### What is a hail swath?

The area a storm's radar hail covered, drawn as bands of hail size. A storm report's map shows the swath around the property and the band the property was in.