Parameter optimization in the detection of bounded weak echo regions in meteorological radar images

The bounded weak echo region (BWER) is a radar signature associated with severe storms. The presence of a BWER within a storm could potentially aid meteorologists in the identification and classification of severe storms. The identification of BWERs is a difficult task as the characteristics are not precisely defined. Meteorologists can identify BWERs within storm cells, but the exact constraints they use are not easy to state. To assist in the identification of BWERs, a method is being developed to automatically process the huge amount of data that must be analyzed in real-time during severe storms. We propose a three step algorithm: (1) detecting candidate regions in 2-D radar sweeps by locating local reflectivity minima, (2) filtering the candidate regions by applying more complex constraints based on BWER characteristic features, and (3) fitting a 3-D model to the likely BWER locations