In Part I (Lan et al., 2010) of this study the feasibility of using an ensemble Kalman filter (EnKF) for storm-scale data assimilation of simulated Doppler radar data is demonstrated assuming a perfect forecast model for a supercell storm．The current study explores the performance of the EnKF in the presence of significant model errors due to microphysical parameterizations．The truth is generated by model simulation using the Lin ice microphysical parameterization．In the first group of EnKF experiments, only one optional microphysical parameterization is used in each experiment to include the model error．The results show that the EnKF performance is seriously degraded, especially for microphysical variables．Then some multi-scheme EnKF experiments are performed．A multi-scheme ensemble forecast that combines all different microphysical parameterization schemes can significantly improve the EnKF performance．The microphysical variables can also be well retrieved．A more targeted multi-scheme is also tested．In this experiment, the multi-scheme ensemble only includes ice microphysical parameterization schemes and the performance of EnKF is further improved.