The role of internal variability and atmospheric teleconnections in determining Arctic sea ice loss

Abstract:

Recent estimates suggest that internal variability may have contributed up to 50% of the observed decline in September Arctic sea-ice extent (SIE) from 1979 to the present. The processes determining these changes are thought to be sensitive to remote sea surface temperature (SST) variability in the subtropical Pacific Ocean, which drives Arctic warming and sea ice melt via an atmospheric teleconnection. However, the short satellite observation record has made it difficult to further examine this relationship. Here, we use 30 fully-coupled global climate models (GCMs) participating in CMIP5 to explore this teleconnection and analyze its stationarity over longer timescales. We find that most GCMs can temporarily simulate the teleconnection in continuous 40-year segments, but exhibit considerable teleconnection variability on multidecadal timescales. No GCM simulates the teleconnection on centennial timescales. During some periods, positive SST anomalies in the Pacific Ocean lead to negative temperatures in the Arctic and positive Arctic SIE anomalies, while during other periods these same GCMs simulate an opposite teleconnection: positive SST anomalies in the Pacific Ocean lead to positive temperatures in the Arctic and negative Arctic SIE anomalies. These findings suggest that the modulation of Arctic sea ice loss by subtropical Pacific Ocean variability is not fixed in time, questioning the role of Pacific Ocean variability for predicting the future evolution of September Arctic SIE. Using a suite of initial-condition large ensembles (LEs) from six fully-coupled GCMs, we further show that internal variability accounts for as much as 50% or as little as 20% of the total uncertainty for projections of Arctic SIE in the next few decades. Our results highlight the importance of quantifying and validating the role of internal variability in projections and predictions of Arctic SIE, particularly at regional scales.