The last few decades have seen a significant decline in Arctic sea ice, generating concerns about both its causes and its longer-term implications. In this paper, we introduce an empirical technique to examine the dynamics of Arctic sea ice extent. Using quantile autoregression, we find that the negative effect of atmospheric CO₂ is stronger in the upper tail of the ice distribution. We also document that Arctic sea ice dynamics have become more unstable over the last three decades, especially during the summer. The rising summer instability occurs across quantiles, indicating that it is due to the joint effects of rising atmospheric CO₂ and nonlinear feedbacks (and not due to outside shocks). While we do not find evidence of “critical slowing”, we see the increasing instability as a cause for concern. We also use the model to predict the evolution of Arctic sea ice extent under alternative CO₂ concentration scenarios.

在过去的几十年中，北极海冰显着减少，引起了人们对其原因及其长期影响的担忧。在本文中，我们引入了一种经验技术来检验北极海冰范围的动力学。使用分位数自回归，我们发现大气CO ₂的负面影响在冰分布的上尾部更强。我们还记录到，在过去的三十年中，特别是在夏季，北极海冰动力学变得更加不稳定。夏季的不稳定性上升发生在各个分位数上，这表明这是由于大气CO ₂上升引起的共同影响和非线性反馈（而不是由于外部冲击）。尽管我们没有发现“严重放缓”的迹象，但我们将不断增加的不稳定视为令人担忧的原因。我们还使用该模型预测了在替代CO ₂浓度情景下北极海冰范围的演变。