《Table 1.Details of the models with sucient availability of sea ice and atmospheric data.Models in b

  提示：宽带有限、当前游客访问压缩模式

本系列图表出处文件名：随高清版一同展现

《Atmospheric Precursors of and Response to Anomalous Arctic Sea Ice in CMIP5 Models》

1. 获取 高清版本忘记账户？点击这里登录

1. 下载图表忘记账户？点击这里登录

The data used in this study are from the CMIP5 archive.Monthly means from the pre-industrial control simulations are used，as the purpose of the investigation is to examine relationships between sea ice and the atmosphere that occur as part of the natural climate variability.There are 34 models（Table 1）that have the required data available.We first examine this group as a whole，before then using a subset of models with different model genealogy（Knutti et al.，2013），which can be considered to be roughly independent of one another.The model subset is denoted with bold text in Table 1，and is selected such that one model per family is chosen and，where possible，similar horizontal resolutions are used.The results from this subset of models are qualitatively similar to those using the full set of models；however，there are quantitative differences in the magnitude of the regressions and their statistical significance.We argue that the non-independence of the models in the full set leads to overconfidence and，therefore，opt to use the smaller set of independent models here.Selected results from a similarly constructed set of high-top models are also presented.These demonstrate that the specific selection of models does not impact the results qualitatively.