Response of the East Asian summer monsoon to doubled atmospheric CO_2: Coupled climate model simulations and projections under IPCC AR4

摘要

The East Asian (China, Korea and Japan) summer monsoon precipitation and its variability are examined from the outputs of the coupled climate models performing coordinated experiments leading to the Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC AR4). Out of the 22 models examined, 14 reproduce the observed shape of the annual cycle well with peak during the boreal summer (June through August), but with varying magnitude. Three models simulate the maximum a month later and with lower magnitudes. Only one model considerably underestimates the magnitude of the annual cycle. The remaining 4 models show some deviations from the observed. Models are unable to simulate the minimum in July with peaks in June and August associated with northward shifts of the Meiyu-Changma-Baiu precipitation band. The realistic simulation of the annual cycle does not appear to depend on the model resolution. The inter-model variation is slightly larger during summer, implying larger diversity of the models in simulating summer monsoon precipitation. The spatial rainfall patterns are reasonably well simulated by most of the models, with several models able to simulate the precipitation associated with the Meiyu-Changma-Baiu frontal zone and that associated with the location of the subtropical high over the north Pacific. Simulated spatial distribution could be sensitive to model resolution as evidenced by two versions of MIROC3.2 model. The multi-model ensemble (MME) pattern reveals an underestimation of seasonal precipitation over the east coast of China, Korea-Japan peninsular and the adjoining oceanic regions. This may be related with the mass-flux based scheme employed for convective parameterization by majority of the models. Further the inter-model variation of precipitation is about 2 times stronger south of 30° N, than north of this latitude, indicating larger diversity of the coupled models in simulating low latitude precipitation. The simulated inter-annual variability is estimated by computing the mean summer monsoon seasonal rainfall and the coefficient of variability (CV). In general the mean observed seasonal precipitation of 542 mm and CV of 6.7% is very well simulated by most of the models. Except for one model mean seasonal precipitation varies from 400 to 650 mm. However the CV varies from 2 to 9%. Future projections under the radiative forcing of doubled CO_2 scenario are examined for individual models and by the MME technique. Changes in mean precipitation and variability are tested by the t-test and F-ratio respectively to evaluate their statistical significance. The changes in mean precipitation vary from -0.6% (CNRM-CM3) to about 14% (ECHO-G; UKMO-HadCM3). The MME technique reveals an increase varying from 5 to 10%, with an average of 7.8% (greater than the observed CV of 6.7%) over the East Asian region. However the increases are significant over the Korea-Japan peninsula and the adjoining north China region only. The increases may be attributed to the projected intensification of the subtropical high, Meiyu-Changma-Baiu frontal zone and the associated influx of moist air from the Pacific inland. The projected changes in the amount of precipitation are directly proportional to the projected changes in the strength of the subtropical high. Further the MME suggests a possible increase in the length of the summer monsoon precipitation period from late spring through early autumn. The changes in precipitation could be stabilized by controlling the CO_2 emissions.