A Case of Rapid Continental Mesoscale Cyclogenesis. Part I: Model Sensitivity Experiments

Abstract

A rapid mesoscale cyclogenesis event took place over the southeastern United States during 28?29 March 1984. This small-scale cyclone, whose initial radius of circulation was approximately 120 km, was associated with a 3-h pressure fall of 11 mb, rainfall exceeding 60 mm, and numerous tornadoes. The development of this mesoscale cyclone was poorly forecasted by the operational Limited-Area Fine-Mesh Model (LFM). Later experiments with the Nested Grid Model (NGM) and eta model also experienced similar failure. In this paper, the authors present a series of numerical experiments using The Pennsylvania State University?National Center for Atmospheric Research Mesoscale Model version 4 (MM4) with the goat of determining factors crucial to a successful prediction of the surface cyclogenesis. The control experiment simulated the rapid mesoscale cyclogenesis by using a 40-km grid spacing; explicit prediction of cloud water, rainwater, and cloud ice; subgrid cumulus parameterization developed by Grell; and the planetary boundary layer scheme developed by Blackadar. The simulated cyclone track and intensity followed the observations reasonably. The model also produced a precipitation distribution superior to that of the operational forecasts. However, the timing of rapid cyclogenesis and heavy precipitation lagged behind the observations by approximately 6 h. Additional experiments were performed to test the sensitivity of the simulations to latent heat release, precipitation parameterization, surface energy fluxes, horizontal grid resolution, the time of initialization, and the treatment of topography. The authors find that the mesoscale cyclogenesis is the result of interaction between an upper-level disturbance and latent heat release. The occurrence of heavy precipitation is strongly influenced by the supply of warm, moist air in the boundary layer, which is in turn affected by the surface energy fluxes. The treatment of precipitation parameterization and the horizontal grid resolution also exert an influence on the accuracy of the simulation. The mesoscale cyclogenesis is not affected significantly by the Appalachian Mountains during the 24-h simulation period. Because of the diabatic nature of the mesoscale cyclone, this cyclogenesis event is found to be highly sensitive to the quality of initial conditions and, therefore, has limited predictability.