A number of studies have explored the effect of anthropogenic emissions on the development and evolution of precipitation in different types of clouds; however, the magnitude of the effect is still not clear, particularly for the case of deep, mixed-phase clouds. In this study, changes in the parameterization of the autoconversion process were introduced in the Advanced Regional Prediction System (ARPS) model to further evaluate this question. The simulations were initialized with cloud droplet distributions measured from an instrumented C-130 aircraft flying 600-800 km offshore in the intertropical convergence zone during the East Pacific Investigations of Climate (EPIC) project. Two contrasting cases were selected, one with and the other without the influence of anthropogenic aerosols. The simulations indicate that the increased cloud condensation nuclei (CCN) concentrations lead to a delay in the formation of rain and to a decrease in precipitation that reaches the surface as a result of the inhibition of the autoconversion of cloud water to rain water and the subsequent delay in the formation of hail. In addition, hail forms at higher levels in the cloud for the case of anthropogenic CCN. The most important process in the production of rain water in both cases is the melting of hail. A decrease in the mass of hail that falls below the freezing level in the polluted case, leads to a decrease in the resulting precipitation at the surface. Changes in the initial concentration of CCN do not appear to influence the storm strength in terms of updrafts and cloud top height, suggesting little sensitivity of the cloud dynamics. A control case simulation using the old microphysics scheme produces much more precipitation than either of the clean and polluted cases. In addition, the clean case with the modified parameterization shows a better agreement to observations than the control case. It is suggested to use the new scheme to simulate deep convective development over tropical maritime regions.