TIDE LEVEL PREDICTION FOCUSING ON STORM SURGE AND SEICHE BY DYNAMICAL SYSTEMS THEORY

Abstract

 Accurate tide level prediction plays a crucial role in the timely and appropriate operation of coastal protection and port facilities. In addition to storm surges caused by meteorological disturbances such as typhoons and low-pressure systems, seiches resulting from the resonance of sea level fluctuations with the natural period of the bay must be considered, especially in enclosed bays. Even after the passage of a typhoon, there is a risk of flooding damage when seiches coincide with high tide. This study focuses on both storm surge and seiche phenomena in Tokyo Bay and performs data-driven tide level prediction based on dynamical systems theory. The model is trained using features such as sea level pressure, surface wind, observed tide level deviations at multiple locations within the bay, and astronomical tide levels to construct a prediction model for the Tokyo tide observation station. The results demonstrate that the model achieves good prediction accuracy for both storm surge and seiche phenomena, even when considering input errors from weather forecasts, confirming its suitability for practical application.