Spatial Pattern and Zonal Shift of the North Atlantic Oscillation. Part II: Numerical Experiments

Abstract

Abstract In this part, the spatial evolution of an initial dipole anomaly in a prescribed jet is at first investigated by numerically solving linear and nonlinear models without forcing in order to examine how the spatial pattern of a dipole anomaly depends on the meridional distribution of a specified jet. It is shown that in a linear experiment an initial symmetric dipole anomaly in the meridional direction can evolve into a northeast–southwest (NE–SW) or northwest–southeast (NW–SE) tilted dipole structure if the core of this jet is in higher latitudes (the north) or in lower latitudes (the south). This is in agreement with the result predicted by the linear Rossby wave theory in slowly varying media. The conclusion also holds for the nonlinear and unforced experiment. North Atlantic Oscillation (NAO) events are then reproduced in a fully nonlinear barotropic model with a wavemaker that mimics the Atlantic storm-track eddy activity. In the absence of topography the spatial tilting of the eddy-driven NAO pattern is found to be independent of the NAO phase. The eddy-driven NAO pattern for the positive (negative) phase can exhibit a NE–SW (NW–SE) tilting only when the core of a prescribed jet prior to the NAO is confined in the higher latitude (lower latitude) region. However, in the presence of the wavenumber-2 topography (two oceans and continents) in the Northern Hemisphere the spatial tilting of the eddy-driven NAO dipole anomaly can be dependent on the NAO phase. Even when the specified basic flow prior to the NAO is uniform, the eddy-driven positive (negative) NAO phase dipole anomaly can also show a NE–SW (NW–SE) tilting because the northward (southward) shift of the excited westerly jet can occur in the presence of topography. In addition, it is found that when the wavemaker is closer to the position of the initial NAO, the eddy-driven positive (negative) NAO phase pattern can display a whole eastward shift and a more distinct NE–SW (NW–SE) tilting. This thus explains why the first empirical orthogonal function of the NAO pattern observed during 1998–2007 exhibits a more pronounced NE–SW tilting than during 1978–97. It appears that the latitudinal shift of the jet, the large-scale topography, and the zonal position of the Atlantic storm-track eddy activity are three important factors for controlling the spatial tilting and zonal shift of eddy-driven NAO dipole anomalies.