European winter temperature variability in a long coupled model simulation: the contribution of ocean dynamics

• Published in: Climate Dynamics
• Main Authors: Wagner, S., Legutke, S., Zorita, E.
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2005
• Subjects: Greenland; North Atlantic; North Atlantic oscillation
• Online Access: http://hdl.handle.net/11858/00-001M-0000-0011-FEBD-C; http://hdl.handle.net/11858/00-001M-0000-0011-FEBC-E

The interannual variability of the European winter air temperature is partially caused by anomalous atmospheric circulation and the associated advection of air masses, mainly linked to the North Atlantic Oscillation (NAO). However, a considerable part of the temperature variability is not linearly described by atmospheric circulation anomalies. Here, a long control simulation with a coupled atmosphere-ocean climate model is analyzed, with the goal of decomposing the European temperature (ET) anomalies in a part linked to the anomalous atmospheric circulation and a residual. The amount of interannual variability of each contribution is roughly 50%, although at subdecadal (overdecadal) time scales the variability of the residuals is dominant. These residuals are found to be linked to temperature anomalies of the same sign in the whole North Atlantic and Greenland, in contrast to the well-known temperature zonal seesaw associated with the NAO. The association between the residuals and other processes in the North Atlantic has been also analyzed. The thermohaline circulation, closely connected in the model to the intensity of the Gulf Stream, lags the evolution of the temperature residuals by several years and thus is not able to control their evolution. The variability of the oceanic convection in the Northern North Atlantic, on the other hand, correlates with the temperature residual at lags close to zero. It is hypothesized that oceanic convection produces a sea-surface temperature fingerprint that leads to the ET residuals. The implications of these results for multi-year predictability and for empirical climate reconstructions are discussed.