Abstract:Multi-physics ensemble simulation experiments of regional climate in China by CWRF (Climate-Weather Research and Forecasting model) for the period 1979-2016 were applied to analyze and evaluate the simulation of temperature and precipitation over the Tibetan Plateau and its sensitivity to horizontal resolution and parameterization scheme of physical processes.The results show that:1) Compared with global models,CWRF downscaling improves the simulation of temperature and precipitation over the plateau,reduces the root mean square error(RMSE) of the annual temperature cycle by nearly 1℃,and decreases the RMSE of the annual cycle of precipitation by 10-40 mm.At the same time,it significantly improves the correlation coefficient between simulated and measured interannual variability,with the highest increase of 0.6;2) Model's resolution has a significant impact on precipitation simulation.The difference of simulated precipitation by models with the different resolution is as high as 60 mm (54%),and the simulation deviation first decreases and then increases with the increased resolution,and the turning point is about 30 km;3) Physical process parameterization scheme has a significant impact on temperature and precipitation simulation.The difference in monthly average temperature between different schematais 1-4℃,and the difference in monthly precipitation in summer is 20-100 mm.Among them,the radiation scheme has the highest impact on temperature simulation,and the cumulus scheme has the utmost impact on precipitation.This paper provides a basis for the selection of physical parameterization and horizontal resolution for CWRF local optimization.