Multilevel Tower Observations of Vertical Eddy Diffusivity and Mixing Length in the Tropical Cyclone Boundary Layer during Landfalls

• Published in: Journal of the atmospheric sciences 2018-09, Vol.75 (9), p.3159-3168
• Main Authors: Tang, Jie, Zhang, Jun A., Aberson, Sim D., Marks, Frank D., Lei, Xiaotu
• Format: Article
• Language: English
• Subjects: Aircraft; Boundary layers; Cyclones; Diffusion coefficients; Diffusivity; Eddy diffusion; Eddy diffusivity; Estimates; Fetch; Fluxes; Hurricanes; Laboratories; Linear functions; Mathematical models; Methods; Mixing length; Momentum; Momentum flux; Momentum transfer; Numerical analysis; Numerical models; Ocean models; Quality control; Storms; Studies; Surface boundary layer; Tower observations; Towers; Tropical climate; Tropical cyclones; Tropical depressions; Tropical storms; Typhoons; Vortices; Wind; Wind data; Wind direction; Wind profiles; Wind speed
• ISSN: 0022-4928; 1520-0469
• DOI: 10.1175/JAS-D-17-0353.1

This study analyzes the fast-response (20 Hz) wind data collected by a multilevel tower during the landfalls of Tropical Storm Lionrock (1006), Typhoon Fanapi (1011), and Typhoon Megi (1015) in 2010. Turbulent momentum fluxes are calculated using the standard eddy-correlation method. Vertical eddy diffusivity Km and mixing length are estimated using the directly measured momentum fluxes and mean-wind profiles. It is found that the momentum flux increases with wind speed at all four levels. The eddy diffusivity calculated using the direct-flux method is compared to that using a theoretical method in which the vertical eddy diffusivity is formulated as a linear function of the friction velocity and height. It is found that below ~60 m, Km can be approximately parameterized using this theoretical method, though this method overestimates Km for higher altitude, indicating that the surface-layer depth is close to 60 m in the tropical cyclones studied here. It is also found that Km at each level varies with wind direction during landfalls: Km estimated based on observations with landward fetch is significantly larger than that estimated using data with seaward fetch. This result suggests that different parameterizations of Km should be used in the boundary layer schemes of numerical models forecasting tropical cyclones over land versus over the ocean.