George Bryan gave a seminar on “How to make simulated hurricanes look like observed hurricanes”
Among the many uncertainties in numerical and theoretical models of hurricanes, perhaps the greatest uncertainty lies in the parameterizations of boundary-layer turbulence and air-sea exchange. In an attempt to reduce this uncertainty, I have conducted a very large set of idealized simulations of strong hurricanes (Category 4+). The simulations vary only in the specification of surface (10-m) exchange coefficients and/or length scales in a turbulence parameterization. To determine which simulation, among the hundreds conducted, is most realistic, I have examined several simple yet well-observed metrics of hurricane intensity and structure, such as: maximum possible wind; minimum possible pressure; surface wind-pressure relationship; height of maximum wind; and surface inflow angle. An optimal set of values for exchange coefficients and turbulence length scales emerges from this analysis, and these values are reasonably consistent with recent observational and laboratory studies. Using this information, I also explain why Emanuel (1995) concluded that the ratio of surface exchange coefficients for enthalpy and drag (Ce/Cd) is likely about 0.75 in strong hurricanes, and why more recent work suggests this ratio is probably about 0.5.
A video and audio recording of his presentation is available on the anonymous ftp site at: