Spatial patterns of low-frequency sea level pressure (SLP) variability are identified by performing rotated principal component analysis (RPCA) on a long-term (1899–1986) Northern Hemisphere gridded dataset. The analysis is limited to the region 160°E eastward to 40°E due to missing data early in the century. The objective is to identify a comprehensive set of highly recurrent SLP teleconnection patterns; to examine some aspects of their seasonality; and to identify the associated mean winter pressure fields and cyclone frequencies occurring at times of opposite eigenvector polarity. The results are further described in the context of the Southern Oscillation and known midtropospheric teleconnection patterns.
Four low-frequency variability patterns are identified over the Atlantic-European sector, including (i) the North Atlantic 0scillation (NAO), and spatial patterns with SLP variability centers over (ii) the eastern Atlantic (EATL), (iii) southern Europe and the northern Mediterranean basin (SENA). and (iv) Scandinavia (SCAN). The Pacific sector low-frequency variability patterns include the (v) North Pacific Oscillation (NPO), and patterns with centers over (vi) the north-central Pacific (PAC) and (vii) the Bering Sea (BER). The EATL, SENA, SCAN and BER patterns have not been identified in previous SLP RPCA studies. Seasonal variations take place in the location of the primary and secondary centre of SLP variability in each teleconnection.
Each of the Atlantic teleconnections have at least one polarity mode in which cyclones migrate toward Greenland and Iceland. However, the opposite modes in the NAO and EATL are associated with distinct zonally oriented cyclone tracks along latitudes 40°–45°N. One mode of each Pacific pattern is characterized by a zonally oriented cyclone track but the opposite modes are characterized by cyclone maxima in the eastern Pacific (NPO), the western Bering Sea (PAC) and another track that turns northward toward the Bering Sea along the date line (BER). The BER and PAC patterns closely resemble the upper-air west Pacific and Pacific-North American patterns, although the correlations between the RPCA scores and the midtropospheric pattern indices can have small but significant correlations to other midtropospheric patterns. The winter BER scores am significantly lag correlated to the Southern Oscillation index.