Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Tetsu Hara

Second Advisor

Isaac Ginis


Air-sea energy and momentum fluxes constrain the available energy for tropical cyclone intensification. The surface wave field modifies these fluxes in numerous ways. In this study we focus on two sea-state dependent processes. The first is the sea-state dependence of the wind stress over the ocean under tropical cyclone force winds. The second is the sea-state dependence of the upper-ocean turbulence due to Langmuir turbulence, which modifies the mixed-layer deepening and cooling of the ocean surface during the passage of a tropical cyclone. In this study we utilize the WAVEWATCH III (WW3) surface wave model, the one-dimensional General Ocean Turbulence Model (GOTM), and the Princeton Ocean Model (POM) to investigate these processes.

The impact of the surface wave field (sea-state) on the wind stress over the ocean is investigated with fetch-dependent seas under uniform wind and with complex seas under idealized tropical cyclone winds. Two different approaches are employed to calculate the wind stress and the mean wind profile. The near-peak frequency range of the surface wave field is simulated using the WAVEWATCH III model. The high frequency part of the surface wave field is empirically determined using a range of different tail levels. The results suggest that the drag coefficient magnitude is very sensitive to the spectral tail level but is not very sensitive to the drag coefficient calculation methods. The drag coefficients at 40 m/s vary from 1 x 10-3 to 4 x 10-3 depending on the saturation level. The misalignment angle between the wind stress vector and the wind vector is sensitive to the stress calculation method used. In particular, if the cross-wind swell is allowed to contribute to the wind stress, it tends to increase the misalignment angle. Our results predict similar amounts of sea state dependence of the drag coefficient regardless of the approaches taken. More sea-state dependence of the drag coefficient is predicted for tropical cyclones than for aligned growing wind conditions, this enhancement of the drag coefficient may be attributed to swell that is significantly misaligned with local wind. The amount of sea-state dependence of the drag coefficient is sensitive to the tail level and the translation speed of the tropical cyclone.

The upper-ocean turbulence is significantly modified by the Stokes drift of the surface waves because of the Craik-Leibovich vortex force (Langmuir turbulence). Under tropical cyclones the contribution of the surface waves varies significantly depending on complex wind and wave conditions. Therefore, turbulence closure models used in ocean models need to explicitly include the sea-state dependent impacts of the Langmuir turbulence. In this study the K Profile Parameterization (KPP) 1st moment turbulence closure model is modified to include the Langmuir turbulence effect, and its performance is tested against concurrent Large Eddy Simulation (LES) experiments under tropical cyclone conditions. First, the KPP model is tuned to reproduce LES results in conditions of shear turbulence only (KPP-ST). Next, KPP is tuned to typical ocean conditions (with typical Langmuir turbulence) but includes no explicit sea-state dependent modifications (KPP-iLT). Next, the Eulerian currents are replaced by the Lagrangian currents in the KPP equations for calculating the bulk Richardson number and the vertical turbulent momentum flux (KPP-Lag). Finally, an enhancement to the turbulent mixing is introduced as a function of the nondimensional turbulent Langmuir number (KPP-LT). KPP-LT, with the Lagrangian currents replacing the Eulerian currents and the turbulent mixing enhanced, significantly improves the prediction of the upper-ocean temperature and currents compared to the default (unmodified) KPP model under tropical cyclones. This modified KPP model also shows improvements over the default KPP at constant moderate winds (10 m/s).

We then examine differences between simulations with KPP-ST, KPP-iLT, and KPP-LT under tropical cyclones in POM. Both KPP-iLT and KPP-LT enhance sea surface cooling due to vertical mixing compared to KPP-ST, but KPP-iLT significantly underestimates the cooling particularly on the left hand side of propagating storms. KPP-LT significantly reduces and homogenizes currents inside the mixing-layer by enhanced vertical momentum mixing, but KPP-iLT has little impacts on currents. Therefore, KPP-iLT introduces further error in predicting horizontal advection of heat near the cold wake as well as sea surface cooling due to upwelling for stationary and slow moving storms. These results suggest that accurate predictions of the Langmuir turbulence effects on upper-ocean response to a tropical cyclone requires an explicit sea-state dependent Langmuir turbulence parameterization.