Error estimation of spectral parameters for high‐resolution wind and turbulence measurements by wind profiler radars. Issue 12 (15th December 2014)

• Record Type: Journal Article
• Authors: Gan, Tong; Yamamoto, Masayuki K.; Hashiguchi, Hiroyuki; Okamoto, Hajime; Yamamoto, Mamoru
• Is Part Of: Radio science. Volume 49:Issue 12(2014)
• Page Start: 1214
• Page End: 1231
• Publication Date: 2014-12-15
• Subjects: Radio meteorology -- Periodicals; Radio wave propagation -- Periodicals; 621.38405
• DOI: 10.1002/2013RS005369
• Languages: English
• ISSNs: 0048-6604

Abstract

<abstract abstract-type="main" id="rds20199-abs-0001"> <title>Abstract</title> <p id="rds20199-para-0001">Using numerical simulations, we investigated a method for calculating the spectral parameters from Doppler spectra collected by high‐resolution wind profiler radars (WPRs). Because high‐resolution WPRs collect a huge amount of Doppler spectra, calculations must be simple and fast. The proposed method has two steps. In the first step, the echo range (<italic>R</italic><sub>echo</sub>), in which the Doppler spectrum point with peak intensity is contained and all the smoothed Doppler spectrum points have intensities that are greater than the noise intensity, was determined. For producing the smoothed Doppler spectrum, a running average with equal weight (RA) or multitaper method (MTM) was used. In the second step, the spectral parameters were calculated using the Doppler spectrum points within <italic>R</italic><sub>echo</sub>. By comparing the performance of the computation methods using RA and MTM, we concluded that the computation method using RA is more suitable because it has better estimation performance for small spectrum widths and the calculations are faster. Estimation error of the spectral parameters depends on the determination accuracy of the Doppler spectrum peak and <italic>R</italic><sub>echo</sub>. Furthermore, for the case of a 512‐point Doppler spectrum and 13‐point RA, the estimation errors tend to be independent of the signal‐to‐noise ratio (SNR) when the peak level of the Doppler spectrum (<italic>p</italic><sub>est</sub>) is ∼8 dB or more greater than the noise intensity. For <italic>p</italic><sub>est</sub> of <∼8 dB, the estimation errors are well correlated to <italic>p</italic><sub>est</sub> and the SNR. Therefore, the number of incoherent integration times should be determined by considering the SNR and <italic>p</italic><sub>est</sub>.</p> </abstract>