Research on Velocity De-aliasing Method for Ground-based Millimeter wave Cloud Radar

Cloud is an important factor of the atmospheric system. Millimeter wave radar is the new equipment of observing cloud, because millimeter wave radar owns relatively short wavelength which is much more close to the diameters of small particles. It is suitable for detecting non-precipitation cloud and weak precipitation cloud，also detecting the change inside the clouds and three-dimensional structure characteristics of clouds continuously, accurately and quickly. Millimeter wave radar has high sensitivity and resolution to obtain some important clouds parameters, such as radial velocity. Compared to the normal centimeter wave weather radar, millimeter wave radar's wavelength is shorter and the maximum unambiguous velocity is smaller, which more likely causes the velocity aliasing and severely limits the quality of velocity data. Ground-based millimeter wave cloud radar usually uses three different scanning modes: plane position indicator (PPI), range height indicator (RHI) and time height indicator (THI). Lots of research have been done on de-aliasing the velocity aliasing of centimeter wave weather radar's PPI velocity products and these methods are relatively suitable for centimeter wave radar, but research on velocity de-aliasing method for ground-based millimeter wave radar is quite few. With the development and application of the ground-based millimeter wave radar, it's a urgent need to alleviate the problem of cloud radar velocity ambiguity. This research proposes corresponding methods aiming at velocity de-aliasing for the three scanning modes of ground-based millimeter wave radar, involving automated two-dimensional multi-pass velocity de-aliasing algorithm used in PPI, automated two-dimensional continuous extrapolation velocity de-aliasing algorithm used in RHI and THI. Taking into account the noise and data missing includes in the velocity product will influence the effect of velocity de-aliasing, put forward to pre-process the velocity data using noise separation method, k-neighborhood frequency method, fast median filtering method, and interpolation method. Finally, errors of these velocity de-aliasing algorithms are verified and analyzed using the real observed data. As a result, de-aliasing performance of two-dimensional continuous extrapolation method is fine under condition of velocity data being continuous along radial, and the multiple velocity aliasing problem can also be solved. However, it's still a challenge to process velocity aliasing data with jumping spot or area.