Has a warm North Atlantic contributed to recent European cold winters?

The rise of global surface temperature waned during the last decade, despite increasing greenhouse gas concentrations. The temperature changes were most pronounced over northern hemisphere land masses during winter (Cohen et al ). They were largely associated with weakening of the mid-latitude westerly flow. To some, these temperature changes may seem paradoxical in the light of anthropogenic global warming, and thus there is much interest in explaining them. Peings and Magnusdottir ( Environ. Res. Lett. 9 034018) provide evidence that recent warming of the North Atlantic sea surface temperature (SST) may be part of the explanation.

Explanations for the recent cold winters have included the strong reduction in arctic sea ice (Francis and Vavrus 2012) and increases in Eurasian snow cover (Cohen et al 2012). Global warming likely contributed to both and as it is expected to accelerate, we may see more harsh winters. However, cold winters are not uncommon in the historic record (Jones et al 2012), and unforced internal climate variability could also play a major role (Wallace et al 2012).
In particular, North Atlantic SST shows pronounced variability with a time scale of 70-80 years-a phenomenon commonly referred to as Atlantic multidecadal variability (AMV) or the Atlantic Multi-decadal Oscillation. Peings and Magnusdottir (2014 Environ. Res. Lett. 9 034018) suggest that the current warm phase of AMV could also have contributed to recent cold winters. Through analysis of historic observations and atmospheric model experiments they show that the warm phase of AMV increases the occurrence of the negative phase of the North Atlantic Oscillation (NAO), and vice versa for cold phases. Independently, Omrani et al (2014) came to similar conclusions using differently defined indices and a different atmospheric model.
These two studies are exciting because they provide modeling evidence for an extra-tropical atmospheric response to North Atlantic SST in winter. Importantly, the response compares well to observations, not only in terms of its pattern but also its strength. This is in contrast to previous studies that found only a weak atmospheric response (Rodwell et al 1999, Bretherton andBattisti 2000).
Why do these two studies reach new conclusions? One reason might be that previous studies have mostly considered the SST tripole pattern, which is primarily a thermodynamic response to the NAO, while the new studies consider SST patterns associated with AMV. The AMV's basin-wide warming (cooling) is not a direct thermodynamic response to the associated negative (positive) NAO like patterns, and models suggest ocean dynamics are important (Eden and Jung 2001). This point, which Bjerknes (1964) noted, has recently been confirmed Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. through analysis of reconstructed turbulent heat flux data (Gulev et al 2013). Thus, multi-decadal SST variations in the Gulf Stream and its extension may act to drive atmospheric circulation changes on decadal timescales.
However, there have been other studies that investigated the atmospheric response to AMV and these have not found a consistent response in winter (Hodson et al 2010). In both new studies, the extra-tropical heating weakens the meridional temperature gradient and thus reduces baroclinicity (Czaja et al 2003). This leads to less synoptic scale eddy activity and a weakening of the westerly flow. Omrani et al (2014) further show that this leads to an increase in upward propagation of quasi-stationary planetary waves and thus to weakening of stratospheric polar vortex and warming in high-latitude stratosphere. The stratospheric warming propagates in turn down into the troposphere, further reducing the baroclinicity, and thus enhancing the negative NAO-like changes in late winter. This wave driven feedback involving the stratosphere (figure 1) adds to our classical understanding of the NAO-response to extratropical SST (Czaja et al 2003). Omrani et al (2014) stress that poorly representing stratosphere and its variability (e.g., major stratospheric warming) can lead to deficiencies in simulating the wintertime response to the AMV, and this could be a key difference to most previous studies. Peings and Magnusdottir's [Pers. Comm.] model results partly agree with this concept, but the stratospheric response is weaker and less significant, consistent with the model's poorer representation of the stratosphere. Other reasons for differences could be the implementation of sponge-layers at the model top, or the background atmospheric state. For example, the strength and structure of the westerly flow controls the upward propagation of quasi-stationary waves. Thus, further work is required to understand the robustness of these modeling results.
The existence of an extra-tropical response to North Atlantic SST could have far reaching implications. Firstly, it implies the northern hemisphere winter climate may be partly predictable on decadal timescales, as SSTs in this region are predictable (Keenlyside et al 2008). Secondly, it opens the possibility for a coupled ocean-atmosphere mode of variability on multi-decadal timescales, possibly extending predictability further. Lastly, it has implications for interpreting global climate variability (Li et al 2013). The cause of the recent northern hemisphere winter cooling remains an open question. While these two new studies suggest the AMV could have contributed, other recent studies have suggested arctic sea ice reduction, increased Eurasian snow cover, and internal climate variability could all have played a role. Understanding the relative importance of these different factors is a high priority. Coordinated model experiments provide one approach to address this issue. This is a theme of a workshop to be held in Bergen in early June, and to which we hope to interest the international modeling community.