Identification and Separation of Turbulence Echo From the Multipeaked VHF Radar Spectra During Precipitation

Abstract

This paper describes an automated data processing algorithm to identify the turbulence echo in the multipeaked very high-frequency radar spectra, generally observed during the precipitation. Although a few multipeak-picking algorithms exist, they identify the turbulence echo from the clutter and noise, but they do not distinguish turbulence and precipitation echoes. The present algorithm identifies the turbulence echo from multiple peaks based on several criteria, such as the significance of the echoes, separation between the echoes within a sliding window, and a reference turbulence echo. A new method [symmetric method (SM)] is introduced to estimate the moments after proper identification of the turbulence echo, in fact, half portion of the echo. Comparison of the moments estimated by the SM and dual-frequency method (another popular method for the extraction of moments from multipeaked spectra) with those obtained by the routine single-peak detection algorithm reveals large discrepancies, particularly in convective cores and in and above the radar bright band region. In these regions, the error in moments by the traditional single-peak-picking algorithm is found to be substantial. Performance evaluation of these two methods has also been carried out against simulated spectra. The analysis reveals that the error in obtaining moments by the SM and dual-frequency method is less than 15%, except for a few cases. Also, the error in moments by the SM is, in general, smaller than by the dual-frequency method.