Search

The parameterization of physical processes is one of the critical elements to properly simulate the regional climate over eastern China. It is essential to conduct detailed analyses on the effect of physical parameterization schemes on regional climate simulation, to provide more reliable regional climate change information. In this paper, we evaluate the 25-year (1983-2007) summer monsoon climate characteristics of precipitation and surface air temperature by using the regional spectral model (RSM) with different physical schemes. The ensemble results using the reliability ensemble averaging (REA) method are also assessed. The result shows that the RSM model has the capacity to reproduce the spatial patterns, the variations, and the temporal tendency of surface air temperature and precipitation over eastern China. And it tends to predict better climatology characteristics over the Yangtze River basin and the South China. The impact of different physical schemes on RSM simulations is also investigated. Generally, the CLD3 cloud water prediction scheme tends to produce larger precipitation because of its overestimation of the low-level moisture. The systematic biases derived from the KF2 cumulus scheme are larger than those from the RAS scheme. The scale-selective bias correction (SSBC) method improves the simulation of the temporal and spatial characteristics of surface air temperature and precipitation and advances the circulation simulation capacity. The REA ensemble results show significant improvement in simulating temperature and precipitation distribution, which have much higher correlation coefficient and lower root mean square error. The REA result of selected experiments is better than that of nonselected experiments, indicating the necessity of choosing better ensemble samples for ensemble. [ABSTRACT FROM AUTHOR]

This study assesses the performance of spectral nudging methodology in dynamical regional climate downscaling for summer climate over East Asia. The regional climate model NCAR-MM5v3 was used to dynamically downscale the 2.5-degree NCEP/NCAR reanalysis (NNRP) data onto 50-km regional grids. The main focus is the model’s simulation of precipitation. The NCEP/CPC precipitation analysis data were used as the verification. Boreal summers (June, July, and August) in 1991, 1998, and 2003 and heavy floods that occurred in Eastern China were selected for the study. Compared to the control runs (CTLs) without spectral nudging (SN), experiments with SNs greatly reduced systematic errors in upper-level large-scale circulations and were in better agreement with the NNRP. At the same time, SNs outperformed CTLs in simulating model variables near the surface. In comparison with observational precipitation data, spectral nudging also improved the model’s simulation of precipitation in spatial and temporal distributions. SN-simulated precipitation field patterns, including the spatial distribution of monthly mean precipitation band, the seasonal march of major precipitation bands, and the daily variability of regional-averaged time series, show much more consistency with observations than those of the CTL runs. [ABSTRACT FROM AUTHOR]