Improving NEXRAD Velocity Retrievals Using Multi-PRT Scans with Regression Processing

A fundamental data quality issue for pulsed Doppler radars is range–velocity ambiguity, especially at low elevation angles where weather echoes out to greater than 400 km in range need to be detected for effective long-range weather surveillance while also retrieving unambiguous radial velocities of around 30 m s²¹. The U.S. National Weather Service (NWS) Next Generation Weather Radars (NEXRADs) mitigate the range–velocity ambiguities by employing a so-called “split-cut” scan whereby the atmosphere is interrogated twice at the same low elevation angle: once with a surveillance scan and once with a Doppler scan. Due to range overlays in the Doppler scan, the SZ(8/64) systematic phase coding technique is used, but frequently velocities still cannot be resolved, which results in velocity voids (referred to as “purple haze”). There are two new techniques to recover data in the velocity void regions: 1) Doppler velocity recovery and dealiasing (VRAD) and 2) SZ(8/64) regression processing. These two algorithms are combined here for the first time for enhanced Doppler velocity recovery, and it is shown that this combination delivers superior velocity estimates as compared to either algorithm individually. SIGNIFICANCE STATEMENT: Critical to the observation, forecasting, and warnings of storms are the accurate estimation and wide spatial coverage of the Doppler velocity field, especially at low elevation angles where long-range unambiguous reflectivity estimates are required. Next Generation Weather Radar (NEXRAD) employs a uniform pulse repetition time (PRT) Doppler scan where the typical unambiguous range is about 120–150 km. The result is frequent overlaid echoes where the velocity cannot be estimated for each overlay, thus producing voids in velocity data. The presented new velocity recovery algorithm by and large makes it possible to significantly reduce the velocity data voids and thereby potentially reveal critical weather signatures. Having velocity estimates in regions that have been previously unattainable would be a significant and consequential improvement for forecasters and other users of radar data.