Estimating Hail Size from Radar

By: John Miller, P.E.

Weather research is a key ingredient in forensic engineering investigations, but it doesn’t tell the whole story.

Forensic engineering considers all data relevant to a specific scope of work, and weather research is a key ingredient in that data set.  Determining the hail size, direction of impact, and date hail occurred are important in every hail study.  Weather data helps us determine if hail fell at a specific location, the size of the hail, from which direction it fell, and when a hailstorm occurred.

A number of companies produce maps and reports estimating the size and location of hail.  These companies rely on interpreting weather radar to estimate the size of hail that possibly affected a specific location or area.  Weather radar emits rapid bursts of energy into storms.  The returning energy (reflectivity) from rain and hail provide information regarding intensity of the rainfall and the presence of hail.  However, weather radar alone is not capable of categorically discriminating areas of rainfall from areas of hail fall, nor the specific size of the hail.[1]  To determine if hail is present in a given storm, other data must be interpreted.  Companies that provide hail maps and reports use computer programs that rely on a probabilistic approach to determine the presence and size of hail.  Thus, these hail maps and reports rely on the quality of the radar imagery available, the quality of the computer program employed, and the skills of the people who interpret the data.  The quality of the radar imagery depends on date and location.  The further a location is from the radar, the less accurate the images will be.  As radars improve, the accuracy of the data improves.

Hail maps and reports indicate that hail up to a certain size possibly occurred at a given location.  These reports are not evidence that hail actually fell at a particular location on a particular date.  If the probability of each event is not given, and any error rate associated with the data is not provided, the user cannot determine the reliability of the data.  Most, if not all, of these hail reports do not provide probabilities that would indicate how confident they are in the data.  In addition, the reports do not state the methodology used to reach the conclusions or the error rate of the data.

Radar data estimating the location, size, and probability of hail fall is also available from the National Oceanic and Atmospheric Administration’s (NOAA) Severe Weather Data Inventory (SWDI) website.[2]  NOAA provides information on the accuracy of their data, stating, “These NEXRAD Level III products (HI, M, SS, TVS) contain several sources of inaccuracies.  The spatial accuracy of the data decreases with the range from the source NEXRAD radar site… Other factors contribute to the accuracy of these products and algorithms.”  These other factors include fewer “sweeps” of the radar data at further distances which reduces the amount of data available; mountainous terrain interfering with the radar signals; and adaptable parameters which are set by the radar operators which are adjusted for each weather situation and season.  NOAA states, “Variation of these values can cause additional inconsistencies.”[3]  Since the companies that produce hail reports use the NOAA radar data as the basis for their reports, any errors or inconsistencies in the NOAA data are built into these reports.

Ultimately, the best indication of hail size at a specific location is the evidence collected at that location, including clean spots, dents, and abrasions on collateral surfaces as well as damage to components such as asphalt shingles.  Hail reports are useful when trying to determine if and when a hailstorm was present, but they should not be used to determine if hail damage occurred at a specific location.  They are also useful, combined with other NOAA data and site-specific observations, in determining when hail damage occurred.

In forensic engineering, we are often faced with imperfect evidence to determine whether damage occurred, when it occurred, and what caused it.  More data typically results in a better answer.  Relying on a single source, especially one that interprets imperfect data, can lead to the wrong conclusion.

[1] E. DeWayne, et al, “Site-Specific Hailstorm Assessment”, RCI Interface Magazine, Vol. XXVI, No. 2, February 2008
[2] NOAA, NOAA’s Severe Weather Data Inventory, (accessed August 22, 2017).
[3] NOAA, NOAA’s Severe Weather Data Inventory, (accessed August 22, 2017).

About the Expert
John Miller joined Donan in 2009 and currently serves as a forensic engineer, principal consultant, and CAT manager based out of the firm’s Louisville, Kentucky office. His area of expertise is structural engineering with an emphasis on structural damage from impact, wind and flood. Mr. Miller’s project capabilities also include a wide range of civil and structural investigations including hail damage, foundation damage and water intrusion.  View John Miller’s full professional profile here.

Check out other weather-related articles and more in our Knowledge Base!