摘要
本文利用CMAP月降水资料、NCEP再分析资料、NOAA的ERSST资料和日本气象厅海气耦合模式(MRI-CGCM)的输出结果,从东亚夏季风气候态、主模态和年际变率等方面分析了MRI-CGCM模式对东亚夏季风的预测性能,并且利用观测的东亚夏季风指数(EASMI)与模拟PC(principal component)的关系建立多元线性回归方程来订正EASMI(简称PC订正法)。结果表明:MRI-CGCM模式能够较好再现东亚夏季风降水和低层风场的气候态,但模拟的西北太平洋反气旋偏弱、偏东,使得模拟的副热带地区降水量偏小。模式较好地模拟出东亚夏季风降水第一模态(EOF1)及相应的低层风场,能够较好再现出EOF1对应El Ni?o衰减位相;模拟降水的EOF1与观测之间的空间相关系数(ACC)为0.72,且能较好地再现其对应的年际变率,其时间系数PC1与观测之间的相关系数为0.41,能模拟出观测EOF1的2 a和5 a主导周期;但模拟的我国以东梅雨锋区雨带位置偏南,这与模拟的西北太平洋反气旋位置偏南有关。模式对降水第二模态EOF2的模拟能力比EOF1明显下降,模拟EOF2与观测之间的ACC降到0.36;虽然模式能较好地再现出EOF2对应El Ni?o发展位相,但模拟的西太平洋反气旋位置偏南,使得雨带位置偏南,模拟的我国梅雨锋区雨带位于江南,与观测场上江南少雨相反。模式较好地模拟出我国东部夏季降水和气温空间异常分布和年际变化,模拟与观测夏季降水和气温的多年平均ACC分别为0.74和0.68。模式模拟我国东部、江淮流域和华南地区夏季降水多年平均PS评分分别为69、70和68分,略高于我国夏季降水业务预测多年平均评分(65分)。模拟的我国东部夏季气温与观测多年平均PS评分为74分。PC订正后EASMI与实况的相关系数由0.51提高到0.65、符号一致率由84%升到91%、标准差由0.75增大到1.4、大于1个标准差年数由6年变为12年,订正后在模拟变幅偏小和梅雨锋区雨带偏南等方面均有一定的改善,对应西太平洋反气旋位置和梅雨锋区雨带位置与实况较为吻合。
Based on the monthly precipitation data of Climate Prediction Center Merged Analysis of Precipitation（CMAP）, the NCEP reanalysis data, the Extended Reconstruction Sea Surface Temperature（ERSST） from National Oceanic and Atmospheric Administration（NOAA）, and the output of the MRI-CGCM（Meteorological Research Institute Coupled Ocean–Atmosphere General Circulation Model） of Japan Meteorological Agency, the ability of MRI-CGCM to simulate precipitation in the East Asian Summer Monsoon（EASM） region is evaluated from the perspective of the climatology, the primary modes of Empirical Orthogonal Function（EOF） and the interannual variation of the EASM. The multiple linear regression equation is established by the relationship between the observed East Asian Summer Monsoon index（EASMI） and the simulated principal component（PC）, and applied to correct the EASMI（PC correction method）. Results show that the MRI-CGCM can reasonably reproduce the basic EASM rainfall and low-level wind fields. However, the simulated western North Pacific anticyclone is weaker and eastward-shifted than normal, which leads to the underestimation of rainfall over the subtropical region. MRI-CGCM can capture the first leading EOF mode（EOF1） of the EASM rainfall and the corresponding wind fields in lower levels over the EASM region and the decaying phase of El Ni？o. The EOF1 space Correlation Coefficient（ACC） between the simulation and observation is 0.72. The interannual variability of EOF1 is reasonably simulated, and the correlation coefficient between the first component（PC1） of MRI-CGCM simulation and observation is 0.41. The simulated EOF1 well reflects the observed characteristics of EOF1. However, the simulated Mei-yu rainbelt over eastern China shifts southward, which is closely related to the southward shift of the western North Pacific anticyclone. The model ability for the simulation of the second leading EOF mode（EOF2） decreases significantly compared to that for the EOF1. The EOF2 ACC between simulation and observation is 0.36. MRI-CGCM can well reproduce the EOF2 that corresponds to the developing phase of El Ni？o. However, the simulated western North Pacific anticyclone shifts southward abnormally, which leads to the southward shift of the rain belt. The simulated Mei-yu rain belt over China is located in the middle and lower reaches of the Yangtze River, which is contrary to the observation that little rain occurs over this region. The spatial distribution and interannual variability of summer precipitation and temperature anomalies in eastern China are well simulated. The mean ACC between the simulated and observed summer precipitation（temperature） is 0.74（0.68）. The mean predict score（PS） of simulated summer precipitation over eastern China, the Yangtze-Huaihe River valley and southern China are 69, 70, and 68, respectively, which are higher than the average PS（65） of the operational summer precipitation prediction. The mean PS of summer mean temperature in eastern China is 74. Improvements in the PC corrected EASMI are reflected in the correlation coefficient, the anomaly sign consistency rate, the weaker magnitude of the simulated EASMI, and the southward shift of the Mei-yu rain belt. The correlation coefficient between the corrected EASMI and observed EASMI increases from 0.51 to 0.65, the anomaly sign consistency rate changes from 84% to 91%, the standard deviation increases from 0.75 to 1.3, the number of years with greater than one standard deviation changes from 6 to 12, and the locations of the western North Pacific anticyclone and Mei-yu rain belt corresponding to the corrected EASMI are more consistent with observations.
出处
《大气科学》
CSCD
北大核心
2016年第6期1320-1332,共13页
Chinese Journal of Atmospheric Sciences
基金
公益性行业(气象)科研专项201406021
安徽省自然科学基金项目1308085QD69
1408085MD73~~
关键词
东亚夏季风
主模态分析
海气耦合模式
模式订正
East Asian Summer Monsoon（EASM）
EOF
Coupled ocean–atmosphere general circulation model（CGCM）
Model correction
