Extended Abstract (448K)P1.12
The sensitivity of modeled hurricanes to the distribution of vertical sigma levels
F. Carroll Dougherty, Univ. of South Alabama, Mobile, AL; and S. K. Kimball
In the course of studying the development of idealized hurricanes using the Pennsylvania State University/National Center for Atmospheric Research (PSU/NCAR) mesoscale model version 5, MM5, a relationship between vortex stability and vertical resolution became apparent. This has led to a study into the sensitivity of the intensification of a hurricane (as measured by minimum sea-level pressure) to the vertical spacing of the model sigma levels. Unlike the horizontal grid spacing, MM5 allows the vertical grid-spacing of the sigma-levels to vary. Conventionally, higher concentrations of sigma levels are chosen in meteorologically active regions, such as the atmospheric boundary level where the atmosphere and earth's surface meet, and the upper levels where the jet stream occurs. While extensive studies have been performed as to sensitivity of hurricane evolution to horizontal resolution, little has been done regarding the vertical spacing of the sigma values.
A specific case of an idealized hurricane over water at a constant 28C surface temperature developed differently depending on the distribution of the sigma levels, even when the number of sigma levels was held constant. Furthermore, asymmetries developed in the hurricanes over time, leading to the eventual demise of the storms. Using the initial thermodynamic state of the atmosphere, a theoretical potential intensity was calculated for the simulated hurricanes. The value is a minimum central sea-level pressure of 907 mb, or a strong Saffir-Simpson category 5 hurricane. None of the simulations in this study made it beyond the category 4 stage except one where most of the sigma levels were concentrated in the upper atmosphere near the hurricane outflow layer. This storm deepened beyond the theoretical level to 895 mb. A strong outflow layer helps to maintain atmospheric instability and thunderstorm formation in hurricanes, the latter providing the storm's fuel. Therefore, this result seems consistent with hurricane physics. However, hurricanes also need a strong inflow layer to supply warm, moist air to maintain the thunderstorms. It would seem logical, then, that simulations with high concentrations of sigma levels in the boundary layer would also produce intense hurricanes. However, a simulation with a large concentration of sigmas in both the lower and upper levels came closer to its potential intensity than one with a similar concentration in the lower levels, but a low concentration of sigmas in the upper levels. Therefore, it seems that a well-resolved outflow is more important than a well-resolved inflow. These findings and other observations will be presented at the conference.
These results have important implications to hurricane forecasting; models such as MM5 are used operationally to forecast hurricane track and intensity. 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. It is hoped that ultimately a recommendation of an optimal number of levels and their distribution can be provided.
Poster Session 1, Posters
Wednesday, 5 May 2004, 1:30 PM-1:30 PM, Richelieu Room
Previous paper Next paper