The diurnal variations from a high-resolution regional climate model (Regional Spectral Model; RSM) are analyzed from six independent decade long integrations using lateral boundary forcing data from the National Centers for Environmental Prediction Reanalysis 2 (NCEPR2), European Center for Medium-Range Weather Forecasts 40-year Reanalysis and the Twentieth Century Reanalysis (20CR). With each of these lateral boundary forcing data, the RSM is integrated separately using two convection schemes: Relaxed Arakawa-Schubert and Kain-Fritsch. The results show that RSM integrations forced with 20CR have the least fidelity in depicting the seasonal cycle and diurnal variability of precipitation and surface temperature over the Southeastern United States. The remaining four model simulations show comparable skill. The differences in the diurnal amplitude of rainfall during the summer months of the 20CR forced integration from the corresponding NCEPR2 forced integration, for example, is found to be largely from the transient component of the moisture flux convergence. The root mean square error (RMSE) of the seasonal cycle of precipitation and surface temperature of the other four simulations (not forced by 20CR) were comparable to each other and highest in the summer months. But the RMSE of the diurnal amplitude of precipitation and the timing of its diurnal zenith were largest during winter months and least during summer and fall months in the four model simulations (not forced by 20CR). The diurnal amplitude of surface temperature in comparison showed far less fidelity in all models. The phase of the diurnal maximum of surface temperature however showed significantly better validation with corresponding observations in all of the six model simulations.

No abstract

No abstract

No abstract