AMIP模式对大气经向质量输送的模拟评估

从描述南、北半球间大气经向质量传输的角度人手,考察IPCC第4次评估报告提供的8个AMIP大气环流模式对越赤道质量通量输送的模拟性能。结果表明:NCAR、MPI和UKMO模式模拟出的越赤道整层大气质量通量与观测大体相一致;MIROC3模拟的整层大气质量通量年循环与观测结果相去甚远,尤其在夏季模拟出较强的虚假向北大气质量输送;IAP模拟的整层大气质量通量年循环方向与观测结果在7个月份中相反;把垂直大气分为4层.各模式对700 hPa以下(I_1)和300-70 hPa(I_3)两层质量通量的模拟能力普遍较好；对700-300 hPa层(I_2)质量通量模拟结果偏差较大;除MIROC3外,其余模式基本能够模拟出70-10 hPa(I_4)大气质量通量的季节变化.显然,不仅南、北半球间大气存在质量交换,越过其他纬度同样存在着经向大气质量输送,无论冬季、夏季还是年平均,各模式对越过其他纬度(60°S-60°N)经向大气质量输送的模拟结果与观测差异明显。整体权衡，UKMO_HADGEMl在模拟越赤道大气质量通量方面表现突出，MPI_ECHAM5模式优势较明显;NCAR、GISS和GFDL 3个模式在某些压力层内具有较好的模拟水平;MIROC模式对整层、700-300 hPa层的模拟能力较低，而对700 hPa以下层和300-70 hPa层的模拟水平较高；IAP_FGOALS和CNRM模式在模拟越赤道大气质量通量方面存在一定的不足.

Evaluation of AMIP models on simulating atmospher ic mass meridional tranport

By regarding ERA40 reanalysis data which is more authentic as observation, the simulation capabilities of 8 AMIP AGCMs provided by IPCC AR4 (CNRM_CM3, GISS_MODEL_E_R, GFDL_CM2_1, IAP_FGOALS1_0_G, NCAR_ CC-SM3_0, MIROC3_2_MEDRES, MPI_ECHAM5 and UKMO_HADGEM1) were evaluated and compared, which focus on the zonal mean meridional atmospheric mass flux across the equator for the whole and four divided levels. It illuminates that the simula-tions of NCAR, MPI and UKMO are closer to the observation about the whole level atmospheric mass transport. The annual cycle of whole level atmospheric mass flux simulated in MIROC3 presents notable differences with observation. Especially, irreal northern mass flux occurs in summer, IAP has opposite direction of mass flux from observation in 7 months of an annual cycle. Because of the wide horizontal and vertical range, the reasons are uncertain of lacking abilities on modeling atmospheric mass transport across the e-quator for the whole level, but then the lower resolution is one of the reasons besides the dynamics frame and parameterization proce-dures. Dividing the whole atmosphere into four layers by 700, 300 and 70 hPa, the atmospheric mass transport concentrates at P_s -700 hPa and 300 - 70 hPa cross the equator. The 8 models have better simulation abilities at the layers of p_s - 700 hPa (I_1 ) and 300 - 70 hPa ( I_3 ) uniformly, which may probably relates with the easiness of describing atmosphere dynamic process at the lower and upper troposphere, nevertheless there are more differences between the 8 AGCMs at the layer of 700 - 300 hPa ( I_2 ). Except MIROC3, the other models have the abilities of simulating the annual cycle of atmospheric mass transport in 70 - 10 hPa ( I_4 ) ulti-mately. The deficiency of MIROC3 may be as a result of improper management of ozone. Furthermore, not only at the equator,meridional atmospheric mass transportation also occurs at other latitudes. Whether in winter, summer or annual mean, the differences of mass flux among these models are all notable from 60°S to 60°N. The simulation result of winter is better than summer. No matter in summer, winter and annual mean, IAP_ FGOALS1_0_G and NCAR_CCSM3_0 overrate the northern transport between 50°S and 60°S. The all and the one, this study provides not only a fire-new angle of AGCMs evaluation, but also powerful elements while choosing AGCMs for studying atmospheric mass transport and cross-equatorial flow. UKMO_HADGEM1 has good represent whilesimulating atmospheric mass flux across the equator, and MPI _ ECHAM5 show obvious superiority. There are better exhibition of simu-lation on some special levels about NCAR, GISS and GFDL. The simulation ability of MIROC is not good enough on simulating the whole level and 700 - 300 hPa, but good enough on p_s - 700 hPa and 300 - 70 hPa. Some shortages are emerged in IAP_ FGOALS and CNRM model. Therefore, UKMO_ HADGEM1 is the preferred model while studying atmospheric mass meridional transform and those concerned.