Signal processing technique for removal of NUBF induced error in spaceborne Doppler precipitation radar measurements

摘要

A sampling strategy and a signal processing technique are proposed to overcome Non Uniform Beam Filling (NUBF) errors on mean Doppler velocity measurements made by spaceborne weather radars. Effects of non uniformity of rainfall within the main antenna lobe in terms on the accuracy of standard estimators are first briefly shown, so as to point out that the bias introduced by NUBF on mean Doppler velocity estimates can be greater than the standard deviation of the estimated velocity, and that it depends on the along-track distribution of reflectivity. Then the sampling strategy is described, based on an oversampling of the integrated data in the along-track direction in order to retrieve information about the reflectivity pattern at the sub-beam scale. The proposed processing technique, named Combined Frequency-Time (CFT) technique, exploits the time series of spectra at fixed range to resolve the NUBF induced bias. The results and the evaluation of performances achievable by means of CFT, were obtained by applying a 3D spaceborne Doppler radar simulator to a 3D dataset of reflectivity and mean Doppler velocity measured through the NASA/JPL airborne Doppler radar ARMAR. The radar system considered here is a nadir-looking, Ku band radar with a sufficiently wide antenna. It is shown how the error on mean Doppler velocity estimates can be reduced by means of CFT to the level predicted for such a radar system in the case of uniformly filled resolution volume (UBF).