Scholarlink

<p>The estimation of wind has applications in various fields, such as weather forecasting, renewable energy, aviation, fire management, construction, and emergency response in natural disasters. The wind estimation from wind profiling radar signals by using the Doppler beam swinging method is accurate during clear skies. However, wind estimation during rainy conditions suffers from errors associated with Doppler shifts due to falling hydrometeors. The presence of hidden noise and the small amplitude of radar echo signals pose challenges in identifying accurate profiles of Doppler peaks, and when the spectra are contaminated with interference signals, the extraction of the right signal becomes intricate. The present study proposes a hybrid modeling framework integrating the empirical mode decomposition (EMD) to develop a novel algorithm for the detection of exact signals by efficiently removing noise-induced interference. This framework first uses EMD to obtain a set of orthogonal components namely the intrinsic mode functions (IMFs) to remove noise-affected undulating IMFs. Subsequently, the remaining aggregated noise-free signals in all rangebins are subjected to maximum peaks estimation using the EMD-Maxpeak algorithm for capturing the doppler profiles. The wind estimation is performed for the 205&#xa0;MHz VHF radar installed at Cochin (10.04°&#xa0;N, 76.33°&#xa0;E), a tropical coastal station in India. The zonal (u), meridional (v), and vertical (w) winds are estimated, and the efficiency of the new algorithm is tested using collocated and concurrent radiosonde data. This novel method removes external interference leading to improved signal detection and wind estimation, with correlations of 0.52 for u wind and 0.65 for v wind at 95% condifence level. These values outperform the existing algorithm, which shows correlations of −0.13 and 0.15, respectively, for the rainy cases. Additionally, the proposed algorithm achieves lower RMSE values of 5.01 and 2.7 for the zonal and meridional wind components, respectively, compared to 6.3 and 4.2 for the existing algorithm, for the analyzed rainy cases. The new algorithm is a promising tool for denoising wind estimation during rainy conditions which aids in unravelling the complex dynamics of significant weather events such as thunderstorms and monsoon convection.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Empirical mode decomposition for improved radar wind estimation during rainy conditions

  • 期刊论文

  • 发表日期：

    2024年12月31日

查看全文
  • Kavya Johny,
  • M. G. Manoj,
  • V. Rakesh,
  • S. Abhilash

摘要

The estimation of wind has applications in various fields, such as weather forecasting, renewable energy, aviation, fire management, construction, and emergency response in natural disasters. The wind estimation from wind profiling radar signals by using the Doppler beam swinging method is accurate during clear skies. However, wind estimation during rainy conditions suffers from errors associated with Doppler shifts due to falling hydrometeors. The presence of hidden noise and the small amplitude of radar echo signals pose challenges in identifying accurate profiles of Doppler peaks, and when the spectra are contaminated with interference signals, the extraction of the right signal becomes intricate. The present study proposes a hybrid modeling framework integrating the empirical mode decomposition (EMD) to develop a novel algorithm for the detection of exact signals by efficiently removing noise-induced interference. This framework first uses EMD to obtain a set of orthogonal components namely the intrinsic mode functions (IMFs) to remove noise-affected undulating IMFs. Subsequently, the remaining aggregated noise-free signals in all rangebins are subjected to maximum peaks estimation using the EMD-Maxpeak algorithm for capturing the doppler profiles. The wind estimation is performed for the 205 MHz VHF radar installed at Cochin (10.04° N, 76.33° E), a tropical coastal station in India. The zonal (u), meridional (v), and vertical (w) winds are estimated, and the efficiency of the new algorithm is tested using collocated and concurrent radiosonde data. This novel method removes external interference leading to improved signal detection and wind estimation, with correlations of 0.52 for u wind and 0.65 for v wind at 95% condifence level. These values outperform the existing algorithm, which shows correlations of −0.13 and 0.15, respectively, for the rainy cases. Additionally, the proposed algorithm achieves lower RMSE values of 5.01 and 2.7 for the zonal and meridional wind components, respectively, compared to 6.3 and 4.2 for the existing algorithm, for the analyzed rainy cases. The new algorithm is a promising tool for denoising wind estimation during rainy conditions which aids in unravelling the complex dynamics of significant weather events such as thunderstorms and monsoon convection.