The sensitivity of hurricane simulations to the distribution of vertical levels

A study completed for an idealized hurricane simulation using MM5, over water, at a constant surface temperature (SST), showed that the storm developed very differently as a function of the vertical spacing of the model sigma levels. The distribution of vertical levels in the inflow, outflow and middle layers of the atmosphere clearly affected the intensity, size, and structure of the storms, causing certain processes to be under- or over-resolved. A strong outflow was found to be necessary for proper storm intensification, while a strong inflow layer did not correspond to an intense storm. In fact, when a strong inflow layer was coupled with a weak outflow layer, a particularly weak storm was the result. When too few levels were assigned to the outflow layer, convection was confined to low and mid-tropospheric levels, and too little latent heating occurred at the middle levels to facilitate the formulation of a strong secondary circulation. However, too few levels in the planetary boundary level would cause a storm to intensify beyond its theoretically calculated maximum potential intensity. This study made use of idealized hurricanes, in quiescent flow, over constant SST. This raised questions about the applicability of the results to real cases where environmental flow and varying SSTs may change the results.

In order to investigate the impacts of varying SST and environmental flow, the sensitivity of real case hurricane simulations to the distribution of vertical levels is tested. Using real cases has the additional advantage of being able to validate results against observations. The first case used is Hurricane Danny (1997) as it made landfall at Mobile, Alabama. Two sets of sensitivity studies are included. The first uses a fixed number of sigma levels distributed differently across the atmosphere. These results are compared with the trends identified with the idealized results. A second set of simulations adds more levels in certain regions and tests the sensitivity to an increased vertical resolution. In addition, the sensitivity of different physical parameterizations (boundary layer and micro physics) to vertical level distribution is assessed.

The ultimate goal of these studies is to provide recommendations for an optimal number of sigma levels and their distribution. These recommendations may be sensitive to the choice of physical parameterizations. These results have important implications to hurricane forecasting. If such models are sensitive to the distribution and number of vertical levels, then forecasts may not accurately represent the real situation with serious consequences to life and property.