MULTIYEAR PRESENT-DAY AND 2XCO(2) SIMULATIONS OF MONSOON CLIMATE OVER EASTERN ASIA AND JAPAN WITH A REGIONAL CLIMATE MODEL NESTED IN A GENERAL-CIRCULATION MODEL

Two continuous 5-year-long simulations over eastern Asia and the Japan islands, one for present-day climate (control) and one for climate under doubled carbon dioxide concentration (2xCO(2)) are completed with a regional climate model (RegCM) nested, in a one-way mode, within a general circulation model (GCM). The GCM is run at R15 resolution (4.5x7.5 degrees latitude x longitude), and the RegCM is run at 50-km grid point spacing. In the control run, both the GCM and the RegCM reproduce the seasonal migration of the westerly jet but produce too strong a monsoonal circulation, which results in a significant overestimate of summer precipitation over the eastern Asian continent. The temporal evolution of the eastern Asia summer monsoon, with steady phases separated by more rapid transitions, is reproduced, but the monsoon rain belt reaches too far north, and the occurrence of tropical storms is underestimated. Regionally averaged surface air temperatures are mostly within 1-3 degrees C of observations. Seasonal precipitation amounts over Japan are within 10-35% of observed ones. The narrow Korea and Japan land masses, which are not captured by the GCM grid, substantially affect the simulated surface precipitation climatology in the RegCM. Under 2xCO(2) forcing, warming in the range of 4-11 degrees C is simulated, greater in winter than in summer and increasing toward high latitudes. The strength of the monsoonal circulation increases in 2xCO(2) conditions, leading to a general increase in precipitation over all regions by 10-30%. The simulated precipitation change shows significant regional and, within Japan, subregional structure. The Japan and Korea land masses substantially affect the summer precipitation changes, indicating the need to capture them in numerical simulations of climate Change. Because of the uncertainties in the control simulations, our 2xCO(2) results are intended not to provide climate change scenarios for impact assessments, but only to illustrate the model sensitivity to different forcings. Plans are under way to further test and improve the RegCM performance over the region and to use more recent GCM simulations to drive the RegCM.