Sensitivity of summer precipitation in regional spectral model simulations over eastern China to physical schemes: Daily, extreme and diurnal cycle

摘要

In this study, the capability of the National Centers for Environmental Prediction (NCEP) Regional Spectral Model (RSM) in reproducing the precipitation in eastern China is evaluated with the simulation of 25-year (1983-2007) summertime climate driven by the NCEP-Department of Energy reanalysis. The impacts of different physical schemes on daily precipitation, diurnal cycle and precipitation extremes are studied through detailed analyses. The results show that the RSM model can reproduce the precipitation characteristics over eastern China, and the results are sensitive to the internal model physics and spectral nudging method. The cloud water prediction scheme of CLD3 (CLDn, the cloud schemes, classified according to the complexity microphysical processes) and the modified version of the Kain-Fritsch scheme (KF2) tend to overestimate the summer precipitation by reproducing higher humidity, while the CLD1 and the Relaxed Arakawa Schubert scheme (RAS) always underestimate it. It is also demonstrated that RSM has the capability to reproduce the spatial distributions of the extreme precipitation intensity and duration. The physical schemes have great impact on the simulations of extreme precipitation over eastern China, especially on regional scales. For the diurnal cycle of precipitation, the responses of the simulated amplitude to the physical schemes differ with regions. The RAS scheme underestimates the amplitudes over southeastern China, while the KF2 generates positive biases. The involving of the scale selective bias correction (SSBC) method can properly improve the simulated mean precipitation and diurnal cycle of summer precipitation. However, it shows lower skill in improving extreme precipitation indices over eastern China, especially in the experiments with CLD3 scheme. In summary, the selection of CLD1 and KF2 scheme, involving SSBC method performs better than the other experiments.