The Simple Rectification to Cartesian Space of Folded Radial Velocities from Doppler Radar Sampling

Abstract

Periodic sampling of the Doppler radar return signal at the pulse repetition frequency causes measured velocities to be ambiguous (folded) when true meteorological velocities along the radial direction exceed the Nyquist or folding value. Furthermore, mean radial velocity estimates become more uncertain as the spatial variability of velocity increases or the returned signal strength decreases. These data have conventionally been prepared for such uses as multiple-Doppler radar wind synthesis by unfolding and editing them in the sampling domain (range-azimuth-elevation spherical coordinates). An alternative method of locally (to the output grid point) unfolding the unedited radial velocities during their linear interpolation to a regular Cartesian grid is presented. The method preserves the spatial discontinuities of order twice the Nyquist value associated with velocity folding. A nondimensional velocity quality parameter is also computed which serves to identify interpolated values that contain too much variance to be reliable. Editing of radar data is thereby postponed until all radar data are mapped to the analysis coordinate system. This allows for iterative global unfolding and multiple-Doppler synthesis of complicated convective storm flow patterns. The resolution of folding in such flow fields may require more information than is usually available from single radar radial velocity fields in spherical coordinates. Further, the amount of data that must be subsequently manipulated is reduced about ten-fold in the process of interpolation.