On the sensitivity of seasonal and diurnal precipitation to cumulus parameterization over CORDEX-EA-II

The ability of the Weather Research and Forecasting (WRF) model in simulating the seasonal and diurnal cycles of rainfall over the Coordinated Regional Climate Downscaling Experiment East Asia Phase II (CORDEX-EA-II) domain is validated against the Tropical Rainfall Measuring Mission (TRMM) datasets. A focus is placed on the role of convective parameterization (CP) schemes. A set of numerical experiments at a 25 km resolution for 1998-2009, using six different CPs, is performed to evaluate the physics-dependency of simulation results. All CPs simulate realistic summer mean precipitation and its northward propagation, with the best performance in the Simplified Arakawa-Schubert (SAS). The biases in the seasonal evolution of rainfall are related to the deficiency in simulated low-level winds and the northward propagation of the cyclonic vorticity. The simulated earlier peak time in other CPs is delayed by about 1-2 h by the Kain-Fritsch with a modified trigger function (KFMT), although this scheme shows a disadvantage in the magnitude. The performance of different CPs in simulating diurnal rainfall cycles is dependent on regions, and none of them performs better than the others for all the sub-regions. The initiation of simulated convection is weakly physics-dependent. However, the timing and magnitude of stratiform precipitation differ among the six experiments. A further analysis shows that the dry biases over the lower Yangtze River basin are a result of the weakened southwesterly water vapor transport, while the excessive afternoon rainfall in the Kain-Fritsch (KF) simulation is attributed to the largest positive perturbation in the lower level atmosphere, especially the enhanced vertical transport of humidity.