A revised parameterization method of the subgrid-scale cloud processes in the deep convection scheme has been incorporated in the National Centers for Environmental Prediction (NCEP) Medium-Range Forecast (MRF) model. The revised method of the subgrid-scale processes in the NCEP deep convection scheme, that is, the Simplified Arakawa-Schubert (SAS) scheme include three components; consideration of stochastic cloud processes, parameterization of convective momentum transport, and inclusion of large-scale destabilization effect. Also, a single column model has been developed to examine the effect of the revised SAS scheme. The single column model utilized in this study is based on the NCEP MRF model, and consists of the same physics packages including the revised SAS scheme.

First, to investigate the characteristics of the revised SAS scheme, simulations have been carried out during the 5 days using the single column model forced by the Tropical Ocean Global Atmosphere Coupled Ocean-Atmosphere Response Experiment (TOGA COARE) data. Second, simulations with 5 member ensembles using the NCEP MRF model have been performed for boreal summers of normal year (1996), El Niņo (1997) and La Niņa year (1999) to examine the impact on the seasonal prediction.

The preliminary results show that the revised SAS scheme has a tendency to decrease the amount of convective rainfall slightly. This decrease is mainly due to the representation of the stochastic cloud processes and convective momentum transport parameterization. In the seasonal prediction, the revised method improves the skill of the prediction for precipitation, and this improvement is due to the representation of the well-organized precipitation system over the western Pacific.