since the mid-1950s have been relatively constant. The COADS and Ponta Delgada sea level pressure time series show remarkable agreement: Both indicate that the warming trend was accompanied by a decrease in the strength of the Subtropical High of approximately 4 mb. The pressure time series give the impression that the downward trend in the atmospheric circulation began about 1905 and lasted until 1962 or so, although the steepest continuous decline occurred during 1920 to 1940.
Bjerknes (1959) found a similar pattern of warming along the Gulf Stream during the first quarter of this century (he used differences between 1890 to 1897 and 1925 to 1932 to depict the warming trend). The warming was accompanied by a strengthening of the subtropical anticyclone, in contrast to the results of this study. We have attempted to reproduce the results of Bjerknes (1959) by using data from the COADS. While a warming along the Gulf Stream is hinted at in the COADS, it is much weaker than that depicted in Bjerknes (0.2°C versus 2°C; not shown). The pressure and wind patterns derived from the COADS broadly support Bjerknes's results.
One important mode of variability of the wintertime surface climate over the North Atlantic during this century is characterized by a dipole pattern in SSTs and surface air temperatures, with anomalies of one sign east of Newfoundland, and anomalies of the opposite polarity off the southeast coast of the United States. This pattern has been noted in the post-World War II data by Wallace et al. (1990), Cayan (1992b), and Kushnir (1994). Wind fluctuations occur locally over the regions of large surface-temperature anomalies, with stronger-than-normal winds overlying cooler-than-normal SSTs. The atmospheric-circulation anomalies resemble the West Atlantic pattern. This mode (both for surface temperatures and for surface winds) exhibits variability on quasi-biennial and -decadal time scales. The decadal fluctuations are irregular in length, averaging about 9 years before 1945 and around 12 years afterward. There does not appear to be any difference between the wind-SST relationships on the different time scales.
The local nature of the atmosphere-ocean relationships exhibited by the dipole pattern suggests that surface wind anomalies contribute to the formation of SST anomalies by altering the fluxes of latent and sensible heat at the ocean surface and the strength of vertical mixing in the upper ocean. Support for this interpretation is given by the ocean modeling studies of Alexander (1990) and Luksch et al. (1989).
Further research is needed to understand the origin of the quasi-decadal cycles in the North Atlantic ocean-atmosphere system. One possibility is that 90 years is not long enough to establish the significance of decadal cycles, and those that have occurred may be random in the sense that they are due to internal atmospheric processes. The decadal SST variations could then be interpreted as simply a reflection of low-frequency modulation of the high-frequency (synoptic and monthly) wind forcing. It should be noted that long (100-year) integrations of atmospheric GCMs, when subject to fixed SST boundary conditions, exhibit prominent low-frequency (decadal and longer) fluctuations (James and James, 1980; Feldstein, 1992).
A more inviting possibility is that the quasi-decadal time