Goddard Cumulus Ensemble (GCE) model has been linked with a spectral bin microphysical scheme. Unlike the widely used bulk microphysical scheme which solves mass conservation equations of cloud, rain, ice, and snow, et. al, bin scheme represents each hydrometeor type with explicit mass bins. In the current model, size distributions of ice crystals (column, plate, and dendrite), snow/aggregate, graupel, hail/frozen drop, cloud and rain drops are represented by 33 bins. The stochastic/kinematic equations of each size bin are solved explicitly. Although computationally intensive, bin scheme allows for studies that can not be explicitly performed by a bulk type scheme, e.g., the impact of aerosol and cloud condensation nuclei (CCN) on cloud dynamics.

An idealized tropical mesoscale convective system is simulated using both bulk and spectral bin version of GCE model. The model is initialized with a sounding observed during TOGA COARE experiment. Three sensitivity tests using identical initial and environmental conditions are discussed in this study: simulation using a bulk type microphysical scheme, bin spectral simulation for dirty/polluted air (high CCN concentration), and bin spectral simulation for clean/background air (low CCN concentration). It is shown that variation of cloud CCN has significant impact on storm dynamics. While the dirty air scenario produces similar total rain amount as bulk simulation, the clean air scenario produces much less rain. Convection in clean air scenario is less intense and less organized. There is very little ice involved in the later stages of clean air simulation, whereas ice remains an important part throughout dirty air simulation. The reason of the dichotomous dynamical responses is linked to different cloud droplet sizes, ice microphysics and the cool pool intensity. Changes on average cloud radiation characteristics caused by different aerosol concentrations and implications on cirrus cloud formation will also be discussed.