Accurate wave prediction can help avoid disasters. In this study, the significant wave height (SWH) prediction performances of the recurrent neural network (RNN), long short-term memory network (LSTM), and gated recurrent unit network (GRU) were compared. The 10 m u-component of wind (U10), 10 m v-component of wind (V10), and SWH of the previous 24 h were used as input parameters to predict the SWHs of the future 1, 3, 6, 12, and 24 h. The SWH prediction model was established at three different sites located in the Bohai Sea, the East China Sea, and the South China Sea, separately. The experimental results show that the performance of LSTM and GRU networks based on the gating mechanism was better than that of traditional RNNs, and the performances of the LSTM and GRU networks were comparable. The EMD method was found to be useful in the improvement of the LSTM network to forecast the significant wave heights of 12 and 24 h.

The effect of swell waves on atmospheric boundary layer turbulence under low winds was explored using data from a fixed platform located in the South China Sea. The wind spectra, cospectra, and Ogive curve measured at a height of 8 m above the mean sea surface provided direct evidence that wind stress was affected by swell waves. To interpret such phenomena, an improved approach was derived based on the fact that the total wind stress was the vector sum of turbulent stress and wave‐coherent stress. Different from the approaches of earlier studies, our approach did not align the turbulent stress with the mean wind speed. The influence of swell waves on the magnitude and direction of the total wind stress was analyzed using our approach. The results showed that the wave‐coherent stress derived from our data accounted for 32% of the total wind stress. The magnitude and angle of the wind stress changed by swell waves depended on the relative angle between the turbulent stress and swell direction.

A comparative study was conducted on typhoon intensity factors affecting the marine environment using two representative cases: Typhoon Lekima, which made landfall at Shandong Peninsula, the Northern East China Sea, and Typhoon Muifa, which did not. Using the ADCIRC and SWAN models, we developed a coupled numerical model and applied it to simulate the storm surge and destructive waves caused by typhoons. Three typhoon parameters—maximum wind speed, radius of maximum wind speed, and translation speed—were investigated through sensitivity experiments. The storm surge during the typhoon that made landfall showed a positive correlation with the distance of the typhoon’s center. The maximum significant wave height and storm surge had near-linear growth with a maximum wind speed but decreased with the growth rate of the radius of maximum wind. A rapid typhoon translation speed from 47 km/h to 60 km/h could cause a storm surge resonance phenomenon at the northern coast of the East China Sea.

Wind stress depends on the sea surface roughness, which can be significantly changed by surface wind waves. Based on observations from a fixed platform, we examined the dependences of the sea surface roughness length on dominant wave characteristic parameters (wave age, wave steepness) at moderate wind speeds and under mixed-wave conditions. No obvious trend was found in the wave steepness dependence of sea surface roughness, but a wave steepness threshold behavior was readily identified in the wave age dependence of sea surface roughness. The influence of dominant wind waves on the surface roughness was illustrated using a wind–wave coupling model. The wave steepness threshold behavior is assumed to be related to the onset of dominant wave breaking. The important role of the interaction between swell and wind wave was highlighted, as swell can absorb energy from locally generated wind wave, which subsequently reduces the wave steepness and the probability of dominant wave breaking.