Diagnosis of the Radiation Fluxes in Grapes-GFS and Related Improvements

Grapes-GFS is the global numerical weather prediction model developed by Numerical Weather Prediction Center (NWPC), CMA. The shortwave (SW) radiation scheme is using RRTMG SW (V3.61). In order to diagnose the performance of radiation scheme and its interaction with other physics processes, the CERES monthly mean radiation fluxes are used for comparison with the radiation fluxes calculated by the radiation scheme in Grapes-GFS. These comparisons help us to find the problems in the parameterization of the radiation process and also the input data. Here shows the summer results of our model from June 1^st to August 31^st (JJA) 2013. The seasonal mean (2013, JJA) clear-sky and all-sky shortwave and longwave fluxes of Grapes-GFS are compared with CERES data at the surface and Top of Atmosphere(TOA) upward and downward fluxes respectively. It has been found that the surface downward clear sky shortwave flux is larger than CERES observation, and the surface downward all sky shortwave flux has positive bias too, which spreads all over the north hemisphere. For the reason that aerosol input data for our radiation scheme is zero, a new set of aerosol climatology data is used, which can reduce the positive bias of SW fluxes on the ground. For the seasonal mean, the improvement of surface temperature can be found. Also the sensitive experiment has been done to study the performance of the surface albedo in reducing the biases of SW fluxes at high latitude North hemisphere. The NCEP monthly mean albedo data are used to replace these calculated by the land surface scheme in Grapes-GFS. The results show that the albedo given by land surface model in grapes-GFS especially at South Africa and Greenland may cause large biases in SW radiation flux calculation. By doing these experiments can slightly improve the forecast accuracy of the Grapes model.