92nd American Meteorological Society Annual Meeting (January 22-26, 2012)

Tuesday, 24 January 2012: 9:15 AM
Evaluation of An Improved Storm Cell Identification and Tracking (SCIT) Algorithm Based on DBSCAN Clustering and JPDA Tracking Methods
Room 357 (New Orleans Convention Center )
Jenny Reed, Georgia Tech Research Institute, Atlanta, GA; and J. Trostel

Poster PDF (243.8 kB)

Accurate storm cell identification and tracking is a major challenge in radar and severe weather operations. The storm cell identification and tracking algorithm based on density based spatial clustering with applications in noise (DBSCAN) and joint probabilistic data association (JPDA) has been presented as a solution to this challenging problem. An evaluation of the performance of this algorithm is presented using an objective statistical approach. The storm cell identification and tracking algorithm is applied to a diverse set of reflectivity data cases. The performance of the algorithm is then scored based on a set of bulk statistics. These statistics include median track duration, track linearity, and track vertically integrated liquid (VIL) consistency.

Supplementary URL: