A semi-empirical method for computing storm surges on open coasts during tropical cyclones

Abstract

A new semi-empirical storm surge prediction (SESSP) method is presented that computes tropical cyclone induced storm surge levels as a function of space and time for various storm and geometry input parameters. It distinguishes between five components that contribute to the total surge: the normal, parallel, radial, Ekman, and inverse barometer surge. The parallel surge (caused by shore-parallel gradients in the wind stress) and radial surge (resulting from wind set-up due to the inflow angle of cyclonic wind fields) have not been adequately reported upon before. The empirical relations are derived using the results of thousands of numerical model simulations in which synthetic storms are simulated at coasts with varying cross-shore topography. Non-linear fitting procedures are used to derive surge response functions (SRFs). The SESSP method reproduces peak surge levels within 10 percent of the values computed by the hydrodynamic model. This study showed variable surge response at different types of coast. On steep coasts, surge is dominated by the inverse barometer effect. Along intermediate and mild sloping coasts, the wind set-up has the greatest contribution to the peak surge. The Ekman surge (caused by the Coriolis effect) typically peaks several hours before landfall and can either have a positive or negative contribution to the peak surge. The radial surge becomes relatively more important on mild-sloping coasts. SESSP's computational efficiency allows for the execution of ensemble forecasts with thousands of synthetic storms along large stretches of coast.