Paper Reviewed
Davini, P. and Cagnazzo, C. 2014. On the misinterpretation of the North Atlantic Oscillation in CMIP5 models. Climate Dynamics 43: 1497-1511.

In introducing their study, Davini and Cagnazzo write that the North Atlantic Oscillation or NAO "is the most prominent pattern of regional wintertime variability of the Northern Hemisphere," citing Wallace and Gutzler (1981) and Hurrell et al. (2003); and they go on to state that the NAO "is characterized by a surface pressure dipole between the low pressure mid-latitude system and the high-pressure subtropical system over the North Atlantic," the positive phase of which "identifies a stronger-than-usual meridional gradient of geopotential height over the North Atlantic and the evident split between the subtropical and eddy-driven jet streams, with the latter blowing over Northern Europe," while "the negative phase is associated with a reduced gradient and with the merging of the two jets." More recently, however, and "making use of the idea of teleconnection pattern," they say that "the NAO has been commonly defined as the leading empirical orthogonal function (EOF) of the Euro-Atlantic region," citing Ambaum et al. (2001).

In light of the significance of this phenomenon, the two Italian researchers describe how they "analyzed a group of the CMIP5 models in order to detect which models are able to replicate the correct connection between blocking/cyclonic Rossby wave breaking over Greenland and the negative phase of the NAO proposed by Woollings et al. (2008, 2010)." This they did for the winters of 1951-2005 for 23 models, "comparing their results with the NCEP/NCAR Reanalysis and Twentieth Century Reanalysis," while additionally "making use of several jet and blocking diagnostics."

As a result of these efforts, Davini and Cagnazzo discovered that several state-of-the-art climate models were, as they describe it, "unable to correctly simulate the physical processes connected to the NAO," which they found to be "especially true for models with a strongly underestimated frequency of high-latitude blocking over Greenland." In fact, they report that "in these models the first empirical orthogonal function (EOF1) of the Euro-Atlantic sector can represent at least three different categories of dominant modes of variability associated with different prevalent regions of blocking occurrence and jet stream displacements." And they conclude that "it is therefore possible to show that such 'biased NAOs' are connected with different dynamical processes with respect to the canonical NAO seen in observations."

Enlarging upon these latter observations, Davini and Cagnazzo write that since the models with the largest issues in replicating the NAO dynamical processes "are models with a poleward displaced eddy-driven jet stream, it would be possible that misinterpretations of the EOF-based NAO will be more frequent in a warmer world." More specifically, or "in other words," as they continue, "it would be possible that we will face the rupture of the paradigm that connects the leading EOF of the North Atlantic and the NAO."

References
Ambaum, M., Hoskins, B. and Stephenson, D. 2001. Arctic Oscillation or North Atlantic Oscillation? Journal of Climate 14: 3495-3507.

Hurrell, J., Kushnir, Y., Ottersen, G. and Visbeck, M. 2003. An overview of the North Atlantic Oscillation. Geophysical Monograph Series 134: 1-36.

Wallace, J. and Gutzler, D. 1981. Teleconnections in the geopotential height field during the Northern Hemisphere winter. Monthly Weather Review 109: 784-812.

Woollings, T., Hoskins, B., Blackburn, M. and Berrisford, P. 2008. A new Rossby wave breaking interpretation of the North Atlantic Oscillation. Journal of the Atmospheric Sciences 65: 326-626.

Woollings, T., Hannachi, A., Hoskins, B. and Turner, A. 2010. A regime view of the North Atlantic Oscillation and its response to anthropogenic forcing. Journal of Climate 23: 1291-1307.