CMIP模式对冬季北半球大气环流与青藏高原冬春季气温相关关系的模拟评估

本文评估了36个CMIP5模式和39个CMIP6模式对近期观测中揭示的北半球冬季大气环流与高原冬春气温之间的相关关系的模拟能力。利用最大协方差(MCA)分析方法,计算并比较了观测和模式中冬季北半球200 hPa位势高度场与同后期青藏高原近地面气温的耦合关系。整体而言,大部分CMIP模式能够模拟出显著的冬季北半球大气环流与青藏高原气温之间的相关关系,且CMIP6模式模拟相关特征和作用机制的能力较CMIP5均有所提升。与观测相比,历史情景下36个CMIP5模式中有26个能够模拟出显著的大气环流与同后期高原气温之间的相关关系,其中对于相关的位势高度场空间模态的模拟明显好于对高原气温异常场空间模态的模拟。同情景下39个CMIP6模式中有37个能模拟出显著相关关系,且CMIP6模式更能模拟出观测中MCA模态的位势高度场上北极涛动(AO)和西太平洋遥相关型(WP)反相位叠加的大气环流特征。在对MCA模态时间变率的模拟上,大部分模式都能重现青藏高原整体变暖的趋势,部分模式能够模拟出观测中位势高度场时间主成分的年际变率,并且CMIP6表现要优于CMIP5。对耦合环流型的动力诊断显示,相比CMIP5模式,CMIP6中有更多模式可以模拟出极地—高原之间的遥相关波列,且对波列结构的模拟更完整。

Assessment of CMIP models in simulating the relationship between wintertime atmospheric circulation in the Northern Hemisphere and the winter-spring temperature over the Tibetan Plateau

In this study, the historical runs by 36 CMIP5 models and 39 CMIP6 models were assessed to investigate how CMIP models can simulate such relationship between the winter circulation and spring TP surface temperature. The Maximum Covariance Analysis(MCA) was applied to the NH geopotential height in winter and the surface temperature over Tibetan Plateau(TP) in the following months in model simulations. In general, most CMIP models are capable of simulating the linkage between Northern Hemisphere(NH) circulation and TP temperature, with the performance of CMIP6 models evidently improved. Among the 36 CMIP5 and 39 CMIP6 models, 26 and 37 of them, respectively, can reproduce the significant lead-lag correlations between the circulation and the TP temperature. CMIP6 models show improved performance in capturing the characteristics of circulations, especially in reproducing patterns with an out-of-phase combination of Arctic Oscillation and West Pacific teleconnection pattern in the MCA mode. For temporal variability, almost all models can simulate the warming trend over the TP, but the spatial pattern varies among models. In terms of simulating interannual variability, CMIP6 performs better than CMIP5 as well. The results also show that more CMIP6 models can simulate the Arctic-Tibetan wave trains that favor the occurrence of TP snow in winter, which can further affect the spring surface temperature in the TP.