A Real-Time Four-Dimensional Doppler Dealiasing Scheme

Abstract

A new dealiasing scheme uses the full four-dimensionality available in an operational Doppler radar data stream. It examines one tilt angle at a time, beginning at the highest elevation where clutter is minimal and gate-to-gate shear is typically low compared to the Nyquist velocity. It then dealiases each tilt in descending order until the entire radial velocity volume is corrected. In each tilt, the algorithm performs six simple steps. In the first two steps, a reflectivity threshold and filter are applied to the radial velocity field to remove unwanted noise. The third step initializes dealiasing by attempting to adjust the value of each gate by Nyquist intervals such that it agrees with both the nearest gate in the next higher tilt and the nearest gate in the previous volume. The gates that pass the third step at a high confidence level become the initial values for step four, which consist of correcting the neighboring gates within the scan, while preserving environmental shear as much as possible. In step five, remaining gates are compared to an average of neighboring corrected gates, and anomalous gates are deleted. As a last resort, step six uses a velocity azimuth display (VAD) analysis of the wind field to interpret and correct any remaining isolated echoes. This scheme uses all available data dimensions to interpret and dealias each tilt and is efficient enough to operate on a continuous data stream. It performs reliably even in difficult dealiasing situations and at low Nyquist velocity. During two complex events observed by low-Nyquist Doppler radar in the European Alps, 93% of 4300 tilts were dealiased without error. When errors did occur, they were usually confined to small regions and most frequently resulted from the occurrence of gate-to-gate shear that was impossible to resolve by the Nyquist velocity.