In the current NCEP operational GFS, there are two issues associated with the Zhao and Carr (1997) microphysics scheme.1. A constant critical relative humidity is used in the scheme for the generation and evaporation of cloud water and precipitation. The use of a constant RHcrit is in contradiction with both the observation analysis (Walcek 1994) and the analysis of cloud resolving model simulations by Xu and Randall (1996) and results in problems in generation of cloud condensation and its evaporation and evaporation of rain. 2. A cloud cover based on that of Xu and Randall (1996) is used in the radiation, which is inconsistent with the Sundqvist humidity type of cloud cover used in the microphysics. In order to handle these two issues, we modified the Zhao and Carr (1997, based on Sundqvist 1989) scheme based on an assumed PDF to eliminate the dependence of RHcrit when cloud condensate is present in the grid box. The new scheme is also used in the radiation schemes for cloud cover calculation. As the current cloud scheme underestimates total cloud cover significantly, we will show the impacts of the modified scheme on the total cloud cover, the cloud water path, and the precipitation field.