The performance and evaluation of the damaging downburst prediction and detection algorithm for bow echo storms

Due to the character of the original source materials and the nature of batch digitization, quality control issues may be present in this document. Please report any quality issues you encounter to digital@library.tamu.edu, referencing the URI of the item.Includes bibliographical references (leaves 50-52).Issued also on microfiche from Lange Micrographics.The performance of the Damaging Downburst Prediction and Detection Algorithm (DDPDA) was evaluated for bow echo storm systems using a database of more than 1800 radar volume scans covering more than 150 hours of observations in twelve different storm systems. This report marks the first time that the DDPDA has been evaluated for bow echo storm systems. Moreover, this validation database is one of the largest validation data sets constructed to date. The DDPDA alerts were evaluated using records of observed straight-line wind damage as well as enhanced validation data from detailed storm damage surveys and dual-Doppler wind field retrievals. Moreover, the algorithm was evaluated over both storm scale (10 km radius from the alert) and warning scale (actual county in which alert occurred) to examine the relation to storm structure and impact on operational utility. It was found that on the storm scale the baseline DDPDA produced a 96 percent false alarm rate (FAR) and a 4 percent critical success index (CSI). The probability of detection (POD) was only 32 percent. On a county scale, the POD increased to 61 percent. The FAR was reduced to 89 percent, and the CSI increased to 10 percent. Performance statistics were found to depend significantly on the quality of the validation data. For two of the storm systems with enhanced validation quality from detailed storm damage surveys or dual-Doppler derived wind fields, the FAR decreased to less than 50 percent and the CSI improved to over 30 percent. The same performance statistics were found when the validation was limited to only areas with a high population density. Several individual parameters within the algorithm were systematically modified to determine how well the existing algorithm could perform. These tests did not produce significant improvements in the overall performance statistics of the DDPDA. Based on the storm structures observed in the validation database, suggestions are given for ways to improve the detection and warning of severe straight-line winds in bow echo storm systems