The majority of bulk microphysical schemes that are used in numerical mesoscale models were developed to represent conditions that occur in cold frontal systems or mid-latitude cyclones. It is not known how well these parameterizations represent conditions that are observed in tropical cyclones.

In this paper, existing microphysical parameterization schemes are modified to better represent the microphysical conditions that are observed in tropical cyclones. In particular, new representations of mass-weighted terminal velocities of distributions of snow and graupel particles are developed to account for the range of crystal types that these ice phase species can represent. In addition, new size-dependent distributions for snow, graupel, rain, and cloud ice are incorporated into models to more accurately portray values observed in tropical cyclones. As a result of these new representations, the rate at which snow, graupel, and ice concentrations change due to processes such as aggregation, autoconversion, riming, and accretion may be substantially modified.

Using the Penn State-NCAR mesoscale model MM5, high resolution simulations of Hurricane Erin are conducted covering the time period between September 7 and September 11 2001. Because Erin never achieved landfall, this tropical cyclone represents an ideal environment for better quantifying sensitivities of the storms to the representation of microphysical processes. In this study, we examine how the track, minimum surface pressure, and precipitation amounts vary depending on the scheme used to represent microphysical processes.