The impact of increasing greenhouse gases and Arctic sea ice decline on the subtropical and eddy-driven jet trends

The impact of the interdecadal trends in greenhouse gases (GHGs) and Arctic sea-ice on the Northern Hemisphere subtropical and eddy-driven jets is investigated. The approach that we adopt is based both on the observation that the dominant atmospheric teleconnection patterns have intrinsic time scales of 5 to 10 days and that these patterns tend to form a continuum. We exploit these properties to try to isolate the impact of the individual trends on the recent poleward shift of both jets.

Self-organizing map (SOM) analysis is used to identify the dominant atmospheric teleconnection patterns. Composites analysis of outgoing longwave radiation shows that the occurrence of each SOM patterns is associated with anomalies in convection over the tropical Indian and western Pacific Ocean. In addition, the variability of some SOM patterns is also found to be linked to fluctuations in Arctic sea ice. Linear regression of the winter-mean frequency of occurrence of each SOM pattern reveals that the observed interdecadal poleward trend of both jets can be interpreted as arising from interdecadal changes in the frequency of occurrence of the SOM patterns, which fluctuate on a 6.5 to 7.5 day time scale. In terms of their impact on the jets, GHG driving leads to a poleward shift of both jets. In contrast, the Arctic sea ice decline is associated with a poleward shift of the subtropical jet and an equatorward shift of the eddy-driven jet. As each of the SOM patterns is driven by both high- and low-frequency transient eddy fluxes, the influence of the GHG driving and Arctic sea ice decline on the jets must be manifested through their impact on the eddy fluxes.

The above results suggest that the positive trend in the Northern Annular Mode prior to the early 1990s can be attributed to GHG driving, while the reversal in its trend since then is likely due to the warming effect being overshadowed by the accelerated loss of Arctic sea ice. Statistically significant correlations are found between summer and autumn Arctic sea-ice area and the winter-mean frequency of occurrence of two SOM patterns. This result suggests that seasonal predictability of the circulation, with a lead time of six months, may be possible.