By taking into account the contributions of both locally and remotely generated internal tides, the tidal mixing in the Luzon Strait (LS) and the South China Sea (SCS) is investigated through internal-tide simulation and energetics analysis. A three-dimensional nonhydrostatic high-resolution model driven by four primary tidal constituents (M₂, S₂, K₁, and O₁) is used for the internal-tide simulation. The baroclinic energy budget analysis reveals that the internal tides radiated from the LS are the dominant energy source for the tidal dissipation in the SCS. In the LS, the estimated depth-integrated turbulent kinetic energy dissipation exceeds O(1) W m⁻² atop the two subsurface ridges, with a dissipation rate of >O(10⁻⁷) W kg⁻¹ and diapycnal diffusivity of ~O(10⁻²) m² s⁻¹. In the SCS, the most intense turbulence occurs in the deep-water basin with a dissipation rate of O(10⁻⁸–10⁻⁶) W kg⁻¹ and diapycnal diffusivity of O(10⁻³–10⁻¹) m² s⁻¹ within the ~2000-m water column above the seafloor as well as in the shelfbreak region with a dissipation rate of O(10⁻⁷–10⁻⁶) W kg⁻¹ and diapycnal diffusivity of O(10⁻⁴–10⁻³) m² s⁻¹. These estimated values are consistent with observations reported in previous studies and are at least one order of magnitude larger than those based solely on locally generated internal tides.

Accurate wave height prediction is significant in ports, energy, fisheries, and other offshore operations. In this study, a regional significant wave height prediction model with a high spatial and temporal resolution is proposed based on the ConvLSTM algorithm. The model learns the intrinsic correlations of the data generated by the numerical model, making it possible to combine the correlations between wind and wind waves to improve the predictions. In addition, this study also optimizes the long-term prediction ability of the model through the proposed Mask method and Replace mechanism. The experimental results show that the introduction of the wind field can significantly improve the significant wave height prediction results. The research on the prediction effect of the entire study area and two separate stations shows that the prediction performance of the proposed model is better than the existing methods. The model makes full use of the physical correlation between wind and wind waves, and the validity is up to 24 hours. The 24-hour forecast R² reached 0.69.

Using the “interior + surface quasigeostrophic” (isQG) method, the density and horizontal velocity fields of the ocean's interior can be retrieved from surface data. This method was applied to the Simple Ocean Data Assimilation (SODA) and the Hybrid Coordinate Ocean Model (HYCOM)/Navy Coupled Ocean Data Assimilation (NCODA) reanalysis data sets. The input surface data include sea surface height (SSH), sea surface temperature (SST), sea surface salinity (SSS), and a region‐averaged stratification. The retrieved subsurface fields are compared with reanalysis data for three tested regions, and the results indicate that the isQG method is robust. The isQG method is particularly successful in the energetic regions like the Gulf Stream region with weak stratification, and the Kuroshio region with strong correlation between sea surface density (SSD) and SSH. It also works, though less satisfactorily, in the Agulhas leakage region. The performance of the isQG method in retrieving subsurface fields varies with season, and peaks in winter when the mixed layer is deeper and stratification is weaker. In addition, higher‐resolution data may facilitate the isQG method to achieve a more successful reconstruction for the velocity retrieval. Our results suggested that the isQG method can be used to reconstruct the ocean interior from the satellite‐derived SSH, SST, and SSS data in the near future.

Using the 1985–2013 record of near-surface currents from satellite-tracked drifters, the pseudo-Eulerian statistics of the near-surface circulation in the Indian Ocean (IO) are analyzed. It is found that the distributions of the current velocities and mean kinetic energy (MKE) in the IO are extremely inhomogeneous in space and nonstationary in time. The most energetic regions with climatologic mean velocity over 50 cm s⁻¹ and MKE over 500 cm² s⁻² are found off the eastern coast of Somalia (with maxima of over 100 cm s⁻¹ and 1500 cm² s⁻²) and the equatorial IO, associated with the strong, annually reversing Somalia Current and the twice-a-year eastward equatorial jets. High eddy kinetic energy (EKE) is found in regions of the equatorial IO, western boundary currents, and Agulhas Return Current, with a maximum of over 3000 cm² s⁻² off the eastern coast of Somalia. The lowest EKE (<500 cm² s⁻²) occurs in the south subtropical gyre between 30° and 40°S and the central-eastern Arabian Sea. Annual and semiannual variability is a significant fraction of the total EKE off the eastern coast of Somalia and in the central-eastern equatorial IO. In general, both the MKE and EKE estimated in the present study are qualitatively in agreement with, but quantitatively larger than, estimates from previous studies. These pseudo-Eulerian MKE and EKE fields, based on the most extensive drifter dataset to date, are the most precise in situ estimates to date and can be used to validate satellite and numerical results.

Drag coefficient (C_d) is an essential metric in the calculation of momentum exchange over the air-sea interface and thus has large impacts on the simulation or forecast of the upper ocean state associated with sea surface winds such as storm surges. Generally, C_d is a function of wind speed. However, the exact relationship between C_d and wind speed is still in dispute and the widely-used formula that is a linear function of wind speed in an ocean model could lead to large bias at high wind speed. Here we establish a parabolic model of C_d based on storm surge observations and simulation in the South China Sea (SCS) through a number of tropical cyclone cases. Simulation of storm surges for independent Tropical cyclones (TCs) cases indicates that the new parabolic model of C_d outperforms traditional linear models.