GOODNESS GRACIOUS, GREAT BALLS OF HAIL

by STEPHEN DEL GRECO, NOAA/NESDIS/NCEI
STEVE ANSARI, NOAA/NESDIS/NCEI
MARK PHILLIPS, UNC ASHEVILLE/NEMAC
ED KEARNS, NOAA/NESDIS/NCEI
edited by JEFF CHEN, STAR YING, AND TYRONE GRANDISON, COMMERCE DATA SERVICE
As part of the Commerce Data Usability Project, the NOAA, in collaboration with the Commerce Data Service has created a tutorial that utilizes NOAA's Severe Weather Data Inventory to develop a risk of damaging hail events across the United States. If you have question, feel free to reach out to the Commerce Data Service at DataUsability@doc.gov.


Imagine baseball-sized ice balls falling from the sky.

For some Americans, this is a real weather phenomenon that risks the physical well-being of people and property. In fact, the frozen precipitation resulting from fast updrafts during strong thunder storms can lead to serious damage and harm. Each year, the U.S. sees approximately $1billion in property and crop damage. But how often does it happen? Where do these events normally happen?

As it turns out, the National Oceanic and Atmospheric Administration (NOAA) collects massive amounts of data on precipitation using radar stations across the country. Algorithms can be built on top of the data to track 'hail signatures' detecting when these events occur and the severity. This severe storm data is captured in NOAA's Severe Weather Data Inventory (SWDI) housed within the National Centers for Environmental Information (NCEI).

Finding patterns over time

Between 2005 and 2010, NOAA has detected over 9 million hail signatures events. Often times, individual events on their own do not provide much context of prevailing conditions. In order to contextualize those events, scientists often process data into climatologies, or weather conditions averaged over a period of time. Climatologies can be presented in frequencies or probabilities over various time units (e.g. hourly, daily, monthly) and geographic units (e.g. 1 degree grid cell) for time horizons of 10 to 30 years. The practical applications are many, ranging from mitigating risk of encountering adverse storm events (e.g. when not to go to the beach) to anticipating the best time start growing certain crops. In short, climatologies help set probabilistic expectations.

The following two climatology graphs illustrate the natural rhythm of hail events over months and hours, respectively. A useful climatology typically focuses in on specific geographic areas, but for demonstration purposes, the data is processed on a national level. Each bar represents the proportion of all hail signatures detected in a given time unit over the 10 year period with clear peaks and troughs.

Monthly Climatology (2005 to 2015)

Monthly percent share of hail events


Hourly Climatology (2005 to 2015)

Hourly percent share of hail events


Where are the most heavily impacted areas? By reprocessing 9 million events down into 16,000 equally spaced grid points 0.25 degrees apart (~17 miles), it becomes easier to determine that the Midwest has a far higher chance of experiencing a hail event on any given day. When limiting the data to severe hail events (e.g. hail diameter > 3in), the majority of these events occur in a smaller but significant area of the country.
Hail data is just the tip of the iceberg. NOAA weather data can be processed into climatologies and be applied to determining when to plant a leafy vegetable to when to take a beach vacation.

Mapping the risk over space

Hail Risk (2005 to 2015)

Daily Probability of Hail by 1/4 degree grids (~17 miles). Nearest county to center of grid cell has been geocoded for context.