Real-time hydrometeor classification for the operational forecasting environment

Techniques classifying hydrometeor types from polarimetric radar moments have been well documented within the research community. With operational dual-polarization radars starting to come on-line in various countries and even the U.S television market, it is time to bring these hydrometeor classification techniques to the operational and commercial user. In doing this the algorithm needs to be robust, tunable, and output to a small but important subset of hydrometeor classifications for use in the sometimes frantic operational environment.

We have built upon previous research and developed a fuzzy logic C-band hydrometeor classification algorithm running within the RVP8™ signal processor. The processing power of the RVP8™ allows using the polarimetric moments along one radial for classifying and outputting hydrometeor types ‘real-time' along with the traditional polarimetric moments. The radial approach implies that viable hydrometeor classifications could be available to a forecaster every 20 seconds (3 RPMs) compared to down-stream volume analysis which may have latencies from 5-15 minutes. We believe this is the first implementation of a hydrometeor classification algorithm within a signal processor.