P1.10 A study of oceanic mixing processes: strong wind regime

Monday, 13 June 2005
Thomas Paine B (Hyatt Regency Cambridge, MA)
Li-Wen Wang, National Sun Yat-Sen Univ., Kaohsiung, Taiwan; and H. H. Chia, X. Li, and C. H. Sui

Oceanic mixing processes in a strong wind regime (hurricanes) are studied in this work through 1D- and 3D-model response experiments to surface forcing. We estimate surface momentum and heat fluxes during Hurricane Felix (13-23, August, 1995) from the Bermuda Testbed Mooring (BTM) and ancillary data in the Atlantic. This case is chosen because it represents an open ocean condition. The surface fluxes are used to force two 1D models: a mixed layer model (the Hybrid version developed by D. Chen) and a turbulence closure model (a Mellor-Yamada level 2.5 type, MY2.5). The censor at BTM measured ocean temperature at 25m, 45m, 61m, 70m, 120m, and 150m. It shows a rapid drop of ocean temperature at 25 m and a rapid rise of temperature at 45 m in response to the passage of Hurricane Felix. The warming and cooling rates in the upper 45 m deep surface water are about the same, ~2oC within the first 12 hours, followed by a very slow recovery. This is indicative of a 1D mixing response. Below the surface mixing layer, temperature at 61m and 70 m show a clear inertia oscillation with the amplitude of about 1oC. The simulated evolution of mixed layer by the two 1D models in response to the observed surface fluxes are compared against observed mixed layer. Simulated and observed mixed layer (SST and h) generally have similar variability and magnitude, yet the effect of wave distortion on momentum exchanges at high wind regime (speed larger than 9~10 m/s) is quite significant. The estimate of momentum exchanges for wind speed larger than 25 m/s remains an open problem. In case of Hurricane Felix, the maximum wind speeds measured at BTM is about 25 m/s. In addition to the 1D processes, the strong friction-induced Ekman pumping must be accounted for to explain the total variability of observed mixed layer structure. To estimate the relative contribution by 1D and 3D processes, we use a 3D model. The 3D model is an idealized version of the Princeton Ocean Model. The model domain is 2000 Km in x and y direction (latitude: 10~30 N), 1035m in depth with a horizontal grid length of 25 Km (81x81 grids). The vertical resolution is stretched in 21 layers. We carry out several idealized experiments with an initial temperature profile that contains a 50-m mixed layer of 28oC, a thermocline 50~400m where temperature cools from 28 to 10oC. An initial Rankin vortex (max. wind speed 50 m/s of radius 50 km) is applied at the surface for 12 hours and then off. The analytic solotion for the linear solution can be solved that contains an axisymmetric part and lat-dependent part of inertia oscillations. The simulated inertial pumping, well described by the analytic solution, causes a deeper mixing in the upper thermocline than buoyancy-induced vertical redistribution of heat in the surface layer which is not included in the current experiments. Vortex Rossby waves are induced in and surround the inner core region. More experiments are being performed to include surface heat fluxes and moving hurricane forcing.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner