Analysis and Evaluation of Data-Adaptive Spectral Estimation Algorithms for Processing MST Radar Data

The National Atmospheric Research Laboratory (NARL) situated at Gadanki, Andhra Pradesh, constitutes the primary source for Indian MST radar data (Mesosphere-Stratosphere-Troposphere). The primary purpose of the MST radar is to analyze and explore the atmospheric dynamics. The MST radar is developed with an array antenna in the active phase and consists of 1024 Yagi-Uda antennas being operated at a frequency of 53 MHz. The NARL provides atmospheric wind data by using MST radar. The wind speed parameters are calculated from the radar echo signals by using spectral estimation. In this paper, the analysis and evaluation of data-adaptive spectral estimation algorithms namely nonparametric iterative adaptive approach (IAA) and semiparametric methods—sparse learning via iterative minimization (SLIM) and sparse iterative covariance-based estimation (SPICE)—have been presented. These algorithms exhibit high resolution and low sidelobe levels. The detectability of the radar signals is improved at low signal-to-noise (SNR) conditions. The zonal speed (U), meridional speed (V) and wind speed (W) are computed and are validated using the Global Positioning System (GPS) radiosonde data. From the results, the SPICE algorithm outperforms the other two data-adaptive algorithms and the existing periodogram method. © 2019, Springer Nature Switzerland AG.