A Case Study of Wave Forecast Over South China Sea Using SWAN Model and Ensemble Kalman Filter Method

The marine and atmospheric model’s accuracy mainly depends on the quality of the initial and boundary conditions, numeric, and physics of the model, and on the variation of the phenomena (e.g., chaos of atmosphere) [1-3]. In the wave prediction problem, the wave field is mainly calculated based on the wave–action balance equations (e.g., the WAVEWATCH III, simulating waves nearshore SWAN, wave model WAM) with the given initial and boundary conditions. There are two types of boundary conditions, including surface and open boundaries. The surface boundary is a field of meteorological factors (wind and pressure fields) extracted from meteorological forecasting models. For the open boundaries, the wave direction, period, wave height or wave spectrum from models or observations can be used at a limited area, such as the South China Sea East Sea of Vietnam (SCS). There are two ways to determine the initial condition for the wave model: (i) The initial condition is the state of the calm sea surface until the actual state of the sea surface is reached to continue calculating the forecast (for example, in the SCS, this process usually takes 48 to 72 h); and (ii) using the analysis wave field as the initial condition. This initial field can be either the observational data or the wave field from the previous forecast (the hot start running).