Search

As an extension of previous work in Wang et al. (Ocean Dyn 64:1447-1456, ), this article presents significant waveheights of swell inverted from a 13 month dataset of two high-frequency (HF) phased array radars. As an important intermediate variable in the calculation of significant waveheights, relative swell directions obtained by two different methods from a single radar station are also presented. The impact of the inaccuracy of relative swell direction on the calculation of waveheight is investigated and an alternative way of using constant swell direction is proposed. Radar-inverted swell significant waveheights using different ranges of relative swell directions are investigated. Results are assessed by WAVEWATCH III model hind casts. Analysis of the complete database shows that radar-inverted swell significant waveheights agree reasonably well with model estimates with large scatter. Standard deviation of the difference between the two estimations increases with waveheight, whereas the relative standard deviation, normalized by waveheight, keeps nearly constant. The constant direction scheme of waveheight inversion gives good estimations except for energetic swell exceeding the small perturbation assumption. Statistical analysis suggests that radar measurement uncertainty explains a considerable part of the difference between radar and model estimates. Swell estimates from both radar stations are consistent. This enables combined use of both radar spectra at common radar cells. Use of double spectra solves the ambiguity of relative swell direction, i.e., absolute swell direction is obtained, and effectively improves the accuracy of swell direction by the least-squares method. [ABSTRACT FROM AUTHOR]

This article presents long period ocean wave (swell) frequencies inverted from a 13-month dataset of high-frequency (HF) phased array radars and an assessment of these estimates by comparison with WAVEWATCH III model data. The method of swell frequency inversion from high-frequency radar sea echo Doppler spectra is described. Radar data were collected from a two-site HF Wellen Radar (WERA) radar system on the west coast of Brittany (France) operating at 12 MHz. A standard beam-forming processing technique has been used to obtain Doppler spectra of processed radar cells. Swell frequencies are obtained from the frequencies of particular spectral peaks of the second-order continuum in hourly averaged Doppler spectra. The data coverage of effective Doppler spectra considered for swell frequency estimates shows the influence of islands and shallow water effects. Swell estimates from both radar stations are in good agreement. The comparison of radar-derived results to WAVEWATCH III (WW3) estimates shows that radar measurements agree quite well with model results. The bias and standard deviation between two estimates are very small for swells with frequency less than 0.09 Hz (period >11 s), whereas radar estimates are generally lower than model estimates for shorter swells, along with higher standard deviation. Statistical analysis suggests that radar measurement uncertainty explains most of the difference between radar and model estimates. For each swell event, time series of frequency exhibits a quasi-linear frequency increase which is associated with the dispersive property of wave phase velocity. The use of swell frequency estimates from both radars on common radar cells only slightly increases the accuracy of swell frequency measurement. [ABSTRACT FROM AUTHOR]