耦合气候系统模式FGOALS-s2海洋数据同化试验模拟的冬季Hadley环流长期变化趋势

Simulation of long-term trends in Hadley circulation during boreal winter using an ocean data assimilation scheme with the coupled general circulation model FGOALS-s2

分析了中国科学院大气物理研究所近期气候预测系统DecPreS的初始化试验的模拟结果,该试验本质上是基于耦合气候系统模式FGOALS-s2的海洋数据同化试验.海洋数据同化方案采用了集合最优差值(ensemble optimal interpolation,EnOI)和分析增量更新(incremental analysis update,IAU)相结合的方案.该研究把同化试验与该模式参与第五次"国际耦合模式比较计划"(CMIP5)的历史气候模拟试验(简称历史试验)结果进行了比较,重点研究海温被观测约束的同化试验对NCEP/NCAR再分析资料揭示的北半球冬季Hadley环流增强趋势的再现能力.结果表明,历史试验模拟的北半球冬季Hadley环流在热带地区存在虚假减弱趋势,同化试验有效地改进了上述偏差,模拟的Hadley环流在5°~15°N呈显著增强趋势,与再分析资料更为接近.同化方案有效改进了对北太平洋和赤道东太平洋海温变化趋势的模拟能力,使得模拟的海温梯度变化更趋合理,最终令Hadley环流变化较之历史试验更接近再分析资料.

In the Northern Hemisphere, an intensification of Hadley circulation has been observed during boreal winter since the 1950s. However, state-of-the-art climate models have limited skill in capturing the observed strengthening trend in Hadley circulation, as can be shown by an examination of historical climate simulations derived from several different climate models. Improving the ability of climate models to capture trends in Hadley circulation is a challenging issue facing the climate community. In this study, an improvement in model performance through constraining sea surface temperature（SST） biases was explored using an initialization run from the near-term climate prediction system DecPreS based on the coupled general-circulation model FGOALS-s2 developed by the Institute of Atmospheric Physics（IAP）. The initialization run uses ocean data assimilation with a new assimilation scheme referred to as the Ensemble Optimal Interpolation–Incremental Analysis Update（EnOI-IAU） and applied to the FGOALS-s2 model. The objective of this study is to assess the performance of FGOALS-s2 in reproducing the observed intensification of long-term Hadley circulation during boreal winter when the modeled ocean temperatures are constrained by observational records. By performing a comparative study of the EnOI-IAU run and a conventional historical run of FGOALS-s2 from the Coupled Model Intercomparison Project Phase 5（CMIP5）, it was found that the EnOI-IAU run performed better than the historical run in simulating the observed long-term trend in Hadley circulation during boreal winter. The EnOI-IAU run accurately reproduced the significant strengthening of Hadley circulation in the range 5°–15°N evident in reanalysis data. In contrast, the historical run predicts a significant weakening of Hadley circulation during boreal winter. The improvement in model performance using the EnOI-IAU run is associated with reduced SST biases when the EnOI-IAU scheme is applied to the FGOALS-s2 model. Specifically, compared with the historical run, warm biases in the North Pacific are largely suppressed and areas of warming in the equatorial eastern Pacific are expanded in the EnOI-IAU run. Both the reduced warm biases and increased areas of warming contributed to an increase in SST meridional gradient over the North Pacific in the assimilation run. The increase in SST gradient then leads to a strengthening of Hadley circulation during boreal winter. Therefore, the application of the EnOI-IAU scheme is demonstrated to be an effective method to improve the capability of the FGOALS-s2 model in reproducing long-term trends in Hadley circulation during boreal winter.