Testing and Analysis of storm cell tracking algorithms using established case-study tracks

Using an organized and objective method for testing any centroid-based storm cell tracking algorithm, analysis of the performance of some tracking algorithms is presented. A previous study used simulated tracks for the source of truth data, however, this method is somewhat insufficient due to unrealistic storm cell tracks. In this study, the hand-made track data from several case-study efforts are used for truth data. In addition, the properties of these tracks are analyzed to improve the realism of the track simulator.

A method for statistically analyzing the tracking results is described. Using case studies that represent a diverse set of meteorological scenarios, the results from the trackers are compared and scored using this truth data. The performance of the algorithms in this context are discussed. Limitations of using hand-made track data are also discussed.