资源环境科技发展态势分析平台

Realistic simulation of cloud variability and rainfall by the coupled models still remains a challenge particularly over the Asian Summer Monsoon (ASM). The simulation of the "pool of inhibited cloudiness" (hereafter referred to as PIC) and associated cloud variability have been analysed in the historical run of 26 models which participated in the Coupled Model Intercomparison Project Phase 5 (CMIP5) and it is shown that the current state of the art general circulation models (GCMs) still have difficulties in properly simulating the PIC. The pool covers an area greater than 1 million km(2) between 3 degrees-13 degrees N and 77 degrees-90 degrees E over the southwest Bay of Bengal (BoB); persisting throughout the ASM and interestingly it is surrounded by the deep convective region. The majority of the models could not simulate cloud variability over the ASM. The performance of National Centre for Environmental Prediction (NCEP) Climate Forecast System (CFSv2) version 2.0, which is being used for operational monsoon prediction over the Indian region, is also tested to evaluate its fidelity in capturing PIC. The moist convective process in the default CFSv2 is found to be one of the major sources of uncertainty in its monsoon prediction. In this study, evaluation of the default CFSv2 (CTRL) and suites of modified CFSv2 have been carried out. The default version of CFSv2 has the simplified Arakawa Schubert (SAS) as convection scheme and Zhao and Carr (ZC) as microphysics. In another modification, the SAS is changed to revised simplified Arakawa Schubert (RSAS) keeping the microphysics unchanged. Further, a more physically based cloud scheme (WRF Single Moment 6-class microphysics-WSM6) is used with SAS and RSAS for comparison of simulation of PIC. Among the CMIP5 models, ACESS-1-0, GFDL-CM3, HadGEM2-CC, HadGEM2-ES are able to represent the PIC reasonably well. CFSCR has shown an improved fidelity in comparison to the CTRL, CTRL-WSM and RSAS and other CMIP5 models. The impact of cloud microphysics in CTRL-WSM and CFSCR appears to play an important role in the simulation of the PIC.