Evaluation Of A Single Ice Microphysics And A Gaussian Probability-Density-Function Macrophysics Scheme

Global nonhydrostatic/hydrostatic model of the Global-Regional Integrated forecast System(GRIST)has been developed on the unstructured Voronoi polygon grids that allowing flexible switch between quasi-uniform and variable-resolution modeling.The scale-aware sub-gridscale physical parameterization scheme is important for the variable-resolution GRIST model,in which the moist physical schemes represent the key processes for atmospheric simulations.A single-ice(SI)cloud microphysics scheme and a Gaussian probability density function distribution(Gauss-PDF)statistical cloud scheme were coupled to the SGRSIT single-column model.Two cases in the Tropical Warm Pool International Cloud Experiment(TWP-ICE)and the North American Southern Great Plains Atmospheric Radiation Measurement Experiment 1997(ARM97)were selected to test the moist physical schemes with the SGRIST single-column model.The numerical results were compared with those of the MG08 cloud microphysical and the PBR14 cloud macrophysics schemes in the CAM5 model and the observations.Moist cloud properties were well simulated sucessfully during the convection active period 25 th to30th Jan.,although cloud fraction above 400 h Pa was overeastimated.The SI and Gauss-PDF cloud physics schemes showed reasonable precipitation with appropriate fluctuations and peaks.The ice water content and liquid water content below the melting level became less and more,respectively,especially during the active period of TWP-ICE,with a better agreement with the observations.In the ARM97 experiment,the Gauss-PDF scheme can effectively improve the simulation of low cloud fraction and extreme precipitation.An improved new Gauss-PDF scheme,by modifying the diagnosis of ice stratus,reduced succesfully the cloud fraction in ARM97 and TWP-ICE numerical experiments,which shows better agreement with the observations.The SI scheme was further tested in the 3D GRIST model.Global total simulation of annual mean shows better agreement of the modeled precipitation with the observations in the south of India Ocean、South Asia and high latitudes of open ocean regions.Simulation of annual mean top of the atmosphere cloud radiative effects longwave and flux in desert region of Africa and polar aeras are better agreement with observations because of more high and low cloud were simulated.