Parametric Estimation of Spectral Moments of Overlapped Weather Doppler Echoes by the Use of High-resolution Algorithms

The purpose of this work is the estimation of the spectral moments of Doppler echoes even in the case of strongly overlapped echoes. In such cases, Fourier like techniques provide poor results because of the lack of resolution. Boyer et al.(2001) proposed the MUSIC algorithm for the estimation of the first spectral moment of the echoes and pointed out the very good resolution of this estimator in comparison with Fourier-like techniques. However, this method doesn’t provide the two other spectral moments of interest for meteorological studies (the zeroth and the second moment of the echoes). To fill this lack, we propose the use of Stochastic Maximum Likelihood (SML) for a joint estimation of spectral moments. This method is based on a parametric modelization of the covariance matrix of the time series. The proposed method is validated on the VHF and UHF times series obtained during Thunderstorm observations at the National Astronomy and Ionosphere Center, Arecibo, PR during September and October 1998. The results obtained confirm the great potential of the method. The algorithm is applied on the UHF time series and the step after consists of determining what echo corresponds to the wind and the hydrometeor ones. The reconstructed wind and reflectivity profiles are in a fairly good qualitative agreement with the corresponding profiles obtained by the use of a classical Fourier technique based on both VHF and UHF time series.