Wednesday, 19 August 2009: 4:15 PM
The Canyons (Sheraton Salt Lake City Hotel)
We study the effects of turbulence on hurricane intensity using two versions of a numerical model. In an axisymmetric version of the model, all turbulence must be parameterized; this includes any non-axisymmetric features in hurricanes, such as eyewall mesovortices, boundary-layer roll vortices, vortex Rossby waves, etc. We find that the parameterized turbulence in the axisymmetric model has a dramatic impact: maximum intensity can be changed by a factor of 2 using previously published settings in the turbulence scheme (see Bryan and Rotunno, 2009, MWR). This is because the turbulence parameterization acts to reduce gradients in both angular momentum and moist entropy, which is consistent with weaker intensity (e.g., by consideration of thermal-wind balance). In a three-dimensional version of the model, it is shown that the effects of turbulence must still be parameterized even when the grid spacing is of order 1 km. Only when grid spacing is less than ~100 m can turbulent eddies be resolved explicitly by the modeling system. Analysis of an eddy-resolving simulation (using 62-m grid spacing) shows that turbulence reduces hurricane intensity primarily by reducing the inherent frontogenetical tendency (Emanuel 1997) in the eyewall region. Specifically, the turbulent eddies act to reduce gradients in entropy and angular momentum, leading to lower maximum intensity compared to weak-turbulence simulations (e.g., compared to simulations using 1-km horizontal grid spacing), consistent with results from the axisymmetric model.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner