Forecasting Hurricane Irene -- from Storm Formation until Unexpected Weakening

A number of simulations of Irene was carried out using the same GFS reanalysis data as initial and lateral boundary conditions, but different model physics configurations, with grid spacing of 3 and 1 km. Despite the fact that most model physics configurations allowed one to reproduce general features of Irene's evolution time dependencies of minimum pressure and maximum wind turned out to be very sensitive to microphysical scheme. Vertical cross-sections of radial averages of wind, fluxes, and hydrometeor masses further highlighted model sensitivities to details in cloud microphysics and boundary layer schemes, particularly the production rate of rain water. Further sensitivity tests with Spectral (bin) Microphysics with 3 km grid spacing suggested the possible importance of high aerosol concentrations on storm development, including weakening prior to landfall. When a one-dimensional ocean model was added to the forecast simulations, the forecasts with SBM microphysics best reproduced the observed maximum wind/minimum wind relationships. Overall, these results suggest that there are dynamical processes occurring in Irene's evolution that may not be reproducible by simply varying the standard physics configuration in the ARW model with grid spacings of 3 and 1 km.