UC San Diego

Research that seeks to characterize the relevant features of the ocean response to strong wind forcing, as models evolve towards reality, is instrumental to the development of parameterizations that could be used in future hurricane forecast models. This thesis investigates several aspects of the ocean's response to strong wind forcing. The studies were conducted using a high- resolution ocean model with sub-gridscale parameterizations of shear-instability type vertical mixing and a unique dataset collected during and after the passage of Hurricane Frances to the east of the Caribbean Leeward Islands in 2004. Simulations conducted for a storm moving eastward at a range of constant speeds in various initial ocean environments, showed that the temperature change decreases with increasing storm speed, although the maximum kinetic energy input is generated when the storm residence time is equal to the near-inertial period. This upholds a previous result conducted using a model where vertical mixing depended on wind stress magnitude. The simulated kinematic response to the hurricane included currents at near twice the inertial frequency, which in the mean field were attributed about equally to horizontal, and vertical, advection of horizontal momentum. During the storm passage, the wind stress curl imparted linear and non-linear components to the vertical velocity fields of comparable magnitude. In simulations where a warm or cold core eddy was forced with a uniform wind stress, a vertical velocity circulation was established in the mean field. Expanding on results from previous studies, it is shown that the maximum vertical velocity in this mean field increases approximately linearly with the maximum vorticity in the cyclogeostrophically adjusted current fields in the eddy. Finally, a simulation initialized with in-situ measurements of temperature in the upper 200m, and observation based wind field, was used to choose the drag coefficient parameterization (as a function wind speed) that provides the closest statical match between simulated and observed temperature and near-inertial current fields. Our results argue for a drag coefficient that saturates at 1.5 x10⁻³