Evaluation of a regional spectral model for the East Asian monsoon case studies for July 1987 and 1988

This study examines the National Centers for Environmental Prediction Regio nal Spectral Model (NCEP RSM)'s capability as a tool for providing regional climate details over East Asia during the summertime. Toward this end, per fect boundary condition experiments driven by analysis data are performed. The NCEP-National Center for Atmospheric Research (NCAR) re-analysis is use d to provide large-scale forcings for the RSM configured with an approximat ely 25-km grid over East Asia centered on the Korean peninsula. Month-long simulations are conducted for July, which is the major monsoon precipitatio n period for Korea. The selected years are 1987 and 1988, which correspond to an El Nino and La Nina year, respectively. Overall, the model results are satisfactory in terms of the simulated large -scale features for the different years. Deviation of the simulated large-s cale features from the analysis generally remains small. Domain averaged er ror of the simulated temperature shows a maximum of 0.8K at 300 hPa, and re lative humidity less than 2 % within the entire troposphere. A discernible bias is found in the simulated climate, including a warming in the northern part and a cooling in the southern part of the domain in the lower troposp here. Cooling is dominant in the middle and upper troposphere, with values less than 1K in much of the domain. This deviation commonly appears for bot h years, indicating that the anomaly forecasts relative to a RSM climatolog y can alleviate uncertainties related to a RSM systematic error. The monthl y accumulated precipitation simulated by the RSM compares well with the obs ervations in terms of anomalous precipitation patterns for the two differen t summers. Regional features, including a local maximum and minimum, are no t reproduced well in the RSM. However, the amount of the domain-averaged pr ecipitation for a month is nearly the same as that recorded by observations . The skill of providing the daily variation of simulated precipitation ove r Korea is relatively poor. Evolution of the daily precipitation in the sim ulations generally compares with the observations qualitatively, but the am ount and onset of precipitation events are not well reproduced. The two sev ere weather events causing disastrous rainfall over Korea, a tropical cyclo ne and a mesoscale convective system, are evaluated in detail.