Laboratoire des Siences du climat et de l'Environnement, Orme des Merisiers, bât. 712, Gif-sur-Yvette Cedex 91191, France
Laboratoire des Siences du climat et de l'Environnement, Orme des Merisiers, bât. 712, Gif-sur-Yvette Cedex 91191, France.
Philos Trans A Math Phys Eng Sci. 2015 Nov 13;373(2054). doi: 10.1098/rsta.2014.0424.
Simulations of the climates of the Last Glacial Maximum (LGM), 21 000 years ago, and of the Mid-Holocene (MH), 6000 years ago, allow an analysis of climate feedbacks in climate states that are radically different from today. The analyses of cloud and surface albedo feedbacks show that the shortwave cloud feedback is a major driver of differences between model results. Similar behaviours appear when comparing the LGM and MH simulated changes, highlighting the fingerprint of model physics. Even though the different feedbacks show similarities between the different climate periods, the fact that their relative strength differs from one climate to the other prevents a direct comparison of past and future climate sensitivity. The land-surface feedback also shows large disparities among models even though they all produce positive sea-ice and snow feedbacks. Models have very different sensitivities when considering the vegetation feedback. This feedback has a regional pattern that differs significantly between models and depends on their level of complexity and model biases. Analyses of the MH climate in two versions of the IPSL model provide further indication on the possibilities to assess the role of model biases and model physics on simulated climate changes using past climates for which observations can be used to assess the model results.
对末次冰盛期(LGM),约 21000 年前,和全新世中期(MH),约 6000 年前气候的模拟,使得对与今天大不相同的气候状态下的气候反馈的分析成为可能。对云与地表反照率反馈的分析表明,短波云反馈是模型结果之间差异的主要驱动因素。在比较 LGM 和 MH 模拟的变化时,也出现了类似的行为,突出了模型物理性质的特征。尽管不同的反馈在不同的气候时期表现出相似性,但它们的相对强度在不同气候之间存在差异,这使得过去和未来气候敏感性无法直接进行比较。地表反馈在模型之间也存在很大差异,尽管它们都产生了正的海冰和雪反馈。在考虑植被反馈时,模型的敏感性也有很大差异。这种反馈具有显著的区域模式,不同模型之间存在很大差异,这取决于模型的复杂程度和模型偏差。对 IPSL 模型两个版本的 MH 气候的分析进一步表明,利用有观测数据可供评估模型结果的过去气候,评估模型偏差和模型物理性质对模拟气候变化的作用的可能性。