Simulation of Tropical Cyclones Using Spectral Bin Microphysics

Abstract

1.1 Bulk-parameterization and spectral bin microphysics Latent heat release is the main energetic source of tropical cyclones (TCs). The intensity of TCs depends on the magnitude of latent heat release (convective heating), as well as on its vertical distribution, and spatial pattern. The latent heat release in clouds depends on the rates of microphysical processes such as: condensational growth/evaporation of droplets, deposition/sublimation of ice, riming and freezing/melting. These microphysical processes are closely related to cloud dynamical properties such as vertical velocity, cloud top height, etc.Droplets and the majority of ice crystals arise on aerosol particles (AP) playing the role of cloud condensational nuclei (CCN) and ice nuclei (IN), respectively. It is well known that droplet size distributions (DSD) and precipitation formation depend on the concentration and size of CCN. Therefore, the convective heating in TCs should depend on the properties of AP in the environment of TCs. Most equations describing microphysical processes, including those responsible for cloudaerosol interaction are well known in Cloud Physics. Advanced microphysical schemes are needed in order to describe microphysical processes adequately and to properly take into account microphysical factors (such as aerosols). Yet, advanced microphysical schemes are not generally used in tropical cyclone forecasting models. Rather, until 2006 the current operational TC forecast model developed at the Geophysical Fluid Dynamics Laboratory used large scale convective parameterizations (Kurihara 1973 and a simplified version of the Arakawa and Shubert scheme). Since 2006 this model used a simplified ArakawaSchubert scheme for cumulus parameterization and a simplified version of the Ferrier bulkparameterization for large-scale condensation in cases when supersaturation in grid points is reached (Bender et al 2007). Both schemes are insensitive to aerosols. The development of one and two-moment bulk parameterization schemes and their application in mesoscale models was an important step toward the improvement in the description of convective processes and precipitation in numerical models. These schemes were implemented in mesoscale and/or forecast modles such as the Penn State/NCAR Mesoscale Modeling System Version 5 (MM5) (Dudhia et al., 1997), RAMS (Pielke et al, 1992) and the Weather Research Forecast Model (Skamarock et al, 2005), the operational numerical weather prediction model of the German Weather Service (COSMO) (that is combined with an extended version of the 2-moment bulk scheme by Seifert and Beheng