AN EXAMINATION OF THE HAIL DETECTION
ALGORITHM OVER CENTRAL ALABAMA
Kevin B. Laws, Scott W. Unger, and John Sirmon
National Weather Service, Birmingham Alabama
With the wide variety of decision tools available to forecasters for use during severe weather warning operations, the usefulness and reliability of certain products in the NWS Advanced Weather Interactive Processing System are often questioned. For example, the hail detection algorithm provides the warning meteorologist with maximum estimated hail size, the probability of hail, and the probability of severe hail. This paper evaluated both the reliability of the hail detection algorithm and its usefulness in warning operations.
Estimates of hail size and the probability of hail were compiled from the algorithm and compared to nearly 500 storm reports from the past several years. Initial comparisons of the algorithm-generated severe hail size product and preliminary storm reports showed when the estimated size from the algorithm met severe criteria, defined by the National Weather Service as 19 mm or greater, severe hail was verified 81% of the time. Furthermore, if the estimated size was greater than 40 mm, large hail was verified 87% of the time. Marginally severe hail, defined as 19-40 mm in diameter, had the lowest verification of just 71% of the time.
The vertically integrated liquid and digital vertically integrated liquid products were also evaluated and compared to the hail detection algorithm output and hail size reports. Warning meteorologists have traditionally used VIL along with other environmental variables to create a limited warning decision tool, VIL of the day. This study attempts to add value to this traditional method by comparing the VIL of the day to output from the hail detection algorithm. It is hoped that showing a positive correlation between the algorithm output and actual hail size reports will demonstrate greater usefulness beyond the VIL of the day when issuing severe weather warnings to the general public.
Both a paper and poster were presented at the 88th Annual American Meteorological Society Conference in New Orleans 2008. View the Poster...Here. View the Paper...Here. These files are large and may take a moment to download.