CMIP5模式对AMO与PDO模拟评估及未来预估

Evaluation and future projection of AMO and PDO in CMIP5 models

利用1880—2009年海表温度(Sea Surface Temperature,SST)观测资料以及耦合模式比较计划第五阶段(Coupled Model Intercomparison Project phase 5,CMIP5)中4种情景(piControl、historical、RCP2.6、RCP4.5)下的模拟资料,通过资料对比,评估了CMIP5模式对两个最为重要年代际尺度模态——北大西洋年代际振荡(Atlantic Multidecadal Oscillation,AMO)和太平洋年代际振荡(Pacific Decadal Oscillation,PDO)的模拟性能,并分析了在不同增暖情景下,这两个海洋年代际模态的变化特征。结果表明:在historical和piControl情景下,多模式集合可以再现北太平洋、东太平洋和北大西洋海表温度的年代际变化中心,但模拟的AMO和PDO模态的振幅都偏弱,特别是PDO模态在东太平洋强度的再现能力较弱。与观测资料相比,在historical情景下对AMO和PDO时空特征模拟较好的模式有:CESM1-CAM5、FGOALS-g2、GISS-E2-H-CC、MIROC5和NorESM1-ME,多模式集合则有更好的模拟效果。在不同增暖情景下,AMO与PDO的空间特征基本一致且振幅差随增暖变化不明显,但是伴随全球增暖加强,两模态都呈现方差贡献减小的特征,尤其AMO模态。

Based on the observed sea surface temperature(SST)data from 1880 to 2009 and the simulated SST datasets under four scenarios(piControl,historical,RCP2.6 and RCP4.5)in the CMIP5(Coupled Model Intercomparison Project phase 5)models,this paper evaluated the simulation performance of CMIP5 models for two most important interdecadal scale modes(Atlantic Multidecadal Oscillation(AMO)and Pacific Decadal Oscillation(PDO)),and further investigated their future projection under different warming scenarios.Results show that the multi-model ensemble can reproduce the interdecadal variation centers of sea surface temperature in North Pacific,East Pacific and North Atlantic under historical and piControl scenarios,but the simulatedamplitudes of AMO mode and PDO mode are weak,especially PDO mode in East Pacific.By evaluating the spatial and temporal characteristics of AMO mode and PDO mode under historical scenario,this paper ranks the ability of models that simulate AMO mode and PDO mode,and suggeststhat CESM1-CAM5,FGOALS-g2,GISS-E2-H-CC,MIROC5 and NorESM1-ME are five better models and the multi-model ensemble is best.Under different warming scenarios,the spatial characteristics of AMO mode and PDO mode are basically consistent and the amplitude differencesamong different warming scenarios are not obvious.However,with the increase of global warming,AMO modeand PDOmode are characterized by the decrease of variance contribution,especially AMO mode.