Abstract
Ocean surface current includes drift current, tidal and residual currents, tsunami current and etc. which have various time scales. We have developed a method for decomposing surface currents in Ise Bay obtained by HF ocean radar into tidal, drift and tsunami current component. The model used in the present method includes the multi-AR model with trend and periodic components developed by Kitagawa and Gersch. The trend component extracted by the model, shows residual and drift currents with a longer time scale. The periodic current component with a period of 12 h appears to be S2 and K2 tidal current components. The steady component extracted by a steady AR model, has a time scale of several to several decades of hours and shows tidal current components with a period of longer than 12 h, such as M2, O1 and K1 tide components. It is difficult to extract tsunami current component using surface current data measured at each one hour because tsunami current has a period of several min. to hours. The multi-AR model can detects a tsunami component as a short period noise component, attacking Ise Bay at 18:00 on March 11.
