Wind speed and direction estimation from wave spectra using deep learning

Abstract. High-frequency parts of ocean wave spectra are strongly coupled to the local wind. Measurements can be used estimate sea surface winds. In this study, two deep neural networks (DNNs) were wind speed and direction from first five Fourier coefficients buoys. The DNNs trained by measurements more than 100 meteorological buoys during 2014–2018. It is found that best represent information about 40 min previously because high-frequency portion spectrum integrates preceding conditions. overall root-mean-square error (RMSE) estimated ∼1.1 m s−1, RMSE ∼ 14∘ when 7–25 s−1. This model not only for estimation compact but also quality control