Low-frequency variability in the coupled ocean-atmosphere system at midlatitudes

The winter-mean climate over the North Atlantic ocean shows a low pressure center near Iceland and a large high pressure center ranging from Florida to Spain. Between these centers, an eastward jet prevails, causing cold and dry winter conditions in Northeast America and moderate conditions in Western Europe. On this mean situation, variability shows up on various time scales. The dominant pattern of variability expresses as a weakening and strengthening of the mean pressure centers and is called the North Atlantic Oscillation (NAO). The impact of the NAO on the American and European winter climate is large and the understanding of a possible mechanism which drives the NAO from one phase into another is therefore of great importance. In the North Atlantic climate system, multiple subsystems appear to be highly related to the NAO. For instance, high correlations have been found between specific North Atlantic sea-surface temperature (SST) patterns of variability and the NAO. Therefore, it has often been suggested that the NAO is both an atmospheric and an oceanic phenomenon with possibly a crucial role for ocean-atmosphere coupling. Although the NAO fluctuates on all time scales, there seems to be a preference for low-frequency variability, ranging from a few years to a few decades. In this thesis, we focus on the NAO at the low-frequency time scale. The central question is: How do oceanic and/or atmospheric processes cause low-frequency fluctuations at midlatitudes? In order to explore this question, first, intermediate models are analyzed using a dynamical systems approach, showing the fundamental processes which lead to low-frequency variability at midlatitudes. Then, output of a complex coupled General Circulation Model (GCM) is analyzed, in order to explore whether these fundamental processes can be identified and understood in the more complex climate system. The results of chapters 2 and 3 show two fundamental mechanisms, which might generate low-frequency variability at midlatitudes: nonlinearly induced variability due to rectification processes and/or low-frequency variability arising through an internal ocean mode, the socalled gyre mode. In chapter 4, a robust decadal oscillatory signal is found, both in oceanic and atmospheric fields. In the ocean surface and subsurface fields, anomalies propagate from the northwestern Labrador Sea south-eastward along the coast and bend around Newfoundland before they enter the subtropical gyre. The pressure patterns in the atmosphere show a strong resemblance with those of the North Atlantic Oscillation. A new framework, in which the gyre mode found in chapter 3 is suggested to set the decadal time scale, is presented to explain the decadal variability. The results of chapters 2- 4 are discussed in chapter 5, leading to a general (personal) view on the NAO: the low-frequency variations of the NAO may very well be caused by low-frequency variability in the ocean