Implementation of Adaptive-Bayesian Shrinkage Technique for Obtaining Winds From MST Radar Covering Higher Altitudes

Accurately estimating wind from higher altitudes while using data from mesosphere stratosphere troposphere (MST) radar operating in the very high frequency (VHF) band is challenging. Several techniques implemented previously have their respective advantages and disadvantages. The present study implements an adaptive-Bayesian shrinkage technique with several advantages over its counterparts. The Bayesian mixture model employed implements parametric and nonparametric methods with predictability. It enhances the MST radar return signal and reduces noise optimally and efficiently. It elaborates on the constructed model comprising the likelihood, priors, Bayes rule, and hyperparameters used in the derivation. This technique can detect weak signals at higher altitudes and works better than the implemented denoising-based non-Bayesian techniques, which are contemporary. All cited techniques in this article use uniform signal enhancement, evaluated uniformly using five moments and six quality parameters. The accuracy of the derived wind is validated against the wind obtained from simultaneous GPS radiosonde observations and found to be consistent. The implemented C# code follows the detailed algorithm step-wise and does not use standard routines that make it fit to work in real time. It concludes that the implemented technique can effectively analyze VHF-MST radar signals and retrieve weak signals from higher altitudes ranging from 25.8 to 28.5 km. It enhances the standard 21-km MST radar reference range that is usually obtained using fast Fourier transform (FFT) and is essential for many scientific investigations.