| Doppler weather radar observations play an increasingly important role in numerical weather prediction (NWP) where real-time forecasts of actual storms, initialized by current data, are within reach. A radar network comprised of both S-band and C-band radars in China, called China Next Generation Weather Radar (CINRAD), is one network that could provide the needed observations of quantifying rainfall, typhoon, and hail in near real time. All these applications require substantial automation, adequate data accuracy, and robust quality control (QC) procedures.Two major issues are the removal of noise and clutter, and correcting Doppler radar velocity data that has been aliased. The current NEXRAD Velocity Dealiasing Algorithm (VDA),(Eilts and Smith,1990), is fundamentally based on minimizing velocity gradients along a radial, and has proven to be a robust and reliable algorithm, having been used for a number of years for dealiasing radial velocity measurements and producing operationally useful products in the United States. In this study, we apply the modified NEXRAD algorithms to CINRAD S-band and C-band data for different weather regimes in China,Radar measurements include noisy data, and can be contaminated by clutter, second-trip echoes and side-lobe echoes. It is necessary to remove all noise before beginning dealiasing. The proposed CINRAD S-band algorithm has a method that utilizes range, echo strength and variance of surrounding data to identify noise. The CINRAD S-band algorithm also incorporates an important improvement choosing an initial reference radial near the "zero" line which can reduce the error rate of dealiasing. The algorithm starts from the initial reference radial and goes through each sweep and tilt radial-by-radial in two passes:one in clockwise direction and another in the counterclockwise direction for180°. In the process of dealiasing, gates without correctly dealiased neighbors are marked with a flag. An attempt to dealias these is made by approaching them from the other side. Before replacing the observed value with the dealiased velocity, the dealiased result is compared with the average velocity of all the valid gates in previous four radials and the seven gates nearest to radar. The proposed S-band algorithm was tested for14different weather scenarios consisting of typhoons, squall lines and heavy rains. The proposed algorithm can improve dealiasing results when dealiasing near noisy gates, range folding, missing gates and velocity fields with strong wind shear. The algorithm correctly dealiased more than99%of the aliased radial velocity data when was compared to the manually edited results. And the performance for typhoon and heavy rain cases was slightly better than the results for squall lines. Due to C-band radar’s shorter wavelength (-5cm), the S-band algorithms must be strengthened to address a problem that is approximately twice as difficult. It is necessary to develop robust and reliable algorithms to dealias CINRAD C-band data with the smaller Nyquist interval. The proposed CINRAD C-band algorithm starts from two initial reference radials which are be separated by approximately180°near the zero lines. A linear least-square error check in each radial can help to dealias errors propagated in the azimuthal direction. The velocities for varying azimuths at the same range from the radar should approximately satisfy a quadratic change relationship and can be dealiased with quadratic least-square fitting. Also, the radial velocities along a radius also nearly meet a quadratic change relationship for higher elevation angles (larger than6.0°) when radar observed Typhoon. The proposed C-band algorithm was tested on C-band data collected in China from four observed strong convective weather cases and four synthetic Typhoon cases.The C-band algorithm correctly dealiased about96%of the aliased radial velocity data. And the result is better than the current NEXRAD and CINRAD S-band algorithm when dealiase the C-band data. |
PDF Full Text Request