Two adaptive beamforming techniques are presented for processing the time series data of phased array radars for meteorological applications. The basic concepts of the two techniques are similar to sidelobe canceling and correlation modeling, but their mathematical presentations are new. The first technique is a two-step beamforming process designed for imaging distributed scatterers, such as the atmosphere, to provide improved estimation of power in high gradient reflectivity fields. The process involves employing a non-adaptive spatial transform to obtain a main channel signal in the steered direction, and then subtracting from the main channel signal an adaptively beamformed signal based on the constraint of zero gain in the steered direction and a combined minimum power. With this processing technique, the shape of the mainlobe has been formed using the non-adaptive spatial transform, which results in improved estimation of values in regions with large reflectivity gradients. The second technique presented is a beamforming process designed to exploit the correlation characteristics of weather signals to improve their detectability. Echo signals, as acquired by Doppler weather radars, are an integrated contribution of waves that propagate from many scattering centers in the radar resolution volume. Their contribution could be modeled as a sequence of independent, temporally correlated Gaussian random variables. In contrast, receiver noise is considered white and is not correlated in time. In low signal-to-noise ratio conditions, the autocorrelation at zero lag is strongly biased by noise and hence not ideal to detect signal. Higher lags, 1 and 2, are not theoretically contaminated by the receiver noise, and they are exploited by this technique to provide improved detectability. The two techniques presented are novel tools that are applicable for imaging the atmosphere.

Validation of the techniques are applied using previously collected real data from the Turbulent Eddy Profiler. Preliminary results of the first technique show improved power estimation compared to Capon beamforming, and preliminary results of the second technique show structured features below the threshold previously retrieved using adaptive beamforming alone.