Pronounced climate changes have occurred since the 1970s, including rapid loss of Arctic sea ice, large-scale warming and increased tropical storm activity in the Atlantic. Anthropogenic radiative forcing is likely to have played a major role in these changes, but the relative influence of anthropogenic forcing and natural variability is not well established. The above changes have also occurred during a period in which the North Atlantic Oscillation has shown marked multidecadal variations. Here we investigate the role of the North Atlantic Oscillation in these rapid changes through its influence on the Atlantic meridional overturning circulation and ocean heat transport. We use climate models to show that observed multidecadal variations of the North Atlantic Oscillation can induce multidecadal variations in the Atlantic meridional overturning circulation and poleward ocean heat transport in the Atlantic, extending to the Arctic. Our results suggest that these variations have contributed to the rapid loss of Arctic sea ice, Northern Hemisphere warming, and changing Atlantic tropical storm activity, especially in the late 1990s and early 2000s. These multidecadal variations are superimposed on long-term anthropogenic forcing trends that are the dominant factor in long-term Arctic sea ice loss and hemispheric warming.
Ocean to Climate 
This is a very nice paper with well designed model experiments. These experiments clearly show how AMOC is forced by NAO at decadal to multi-decadal time scales. As discussed in this paper, the main question is where the low frequency NAO signal coming from and if it can be predicted. This age old question is still up in the air.
Rodwell et al. (1999) in their paper appeared in Nature (http://www.nature.com/articles/18648) stated “Our simulations indicate that much of the multiannual to multidecadal variability of the winter NAO over the past half century may be reconstructed from a knowledge of North Atlantic sea surface temperature”. Later, Bretherton and Battisti (2000; http://onlinelibrary.wiley.com/wol1/doi/10.1029/1999GL010910) criticized Rodwell et al. (1999) pointing out that “amplitude of the atmospheric variability is considerably reduced and the air-sea heat fluxes are of the reverse sign to those observed”. They further used a simple linear system to conclude that “the observed SST history in the North Atlantic is primarily driven by a unique and unpredictable realization of dynamical induced atmospheric noise”, invoking Hasselmann’s theory. See latest discussion on this topic by Clement et al. (2015; http://science.sciencemag.org/content/350/6258/320) and Zhang et al. (2016; http://science.sciencemag.org/content/352/6293/1527.1).