scale is a property of the coupled ocean-atmosphere system. We note that decadal and longer time scale SST variations along the boundary between the sub-polar and sub-tropical ocean gyres are obtained in ocean GCMs as a result of self-sustained oscillations of the thermohaline circulation (cf. Weaver and Sarachik, 1991). In addition, atmospheric GCM experiments indicate that the local atmospheric response to an SST anomaly off Newfoundland is such that weak (strong) winds coincide with high (low) SSTs, reinforcing the original SST anomaly (Palmer and Sun, 1985; Lau and Nath, 1990). The implication of these studies is that observed mid-latitude climate anomalies may result from a positive feedback between the atmosphere and the ocean.
A third possibility is that the decadal atmospheric fluctuations are a response to decadal SST variations outside the North Atlantic. This scenario is difficult to test due to the lack of long marine records in the other ocean basins. Our preliminary results indicate that for the three most recent decadal oscillations, SSTs in the North Pacific are not coherent with those in the Atlantic.
One intriguing link we have found is that between sea ice in the Davis Strait-Labrador Sea region and quasi-decadal fluctuations. Figure 10a shows the time series of winter sea-ice concentration anomalies in the Davis Strait/Labrador Sea region from Agnew (1991), based on data from Walsh and Johnson (1979). The circles denote the winter (December-February) anomaly, plotted in the year in which January occurs. The solid curve shows the data smoothed with a 3-point binomial filter. Decadal variability is evident in the sea-ice record, with peaks occurring in the winters of 1957/ 1958, 1971/1972, and 1983/1984. Figure 10b shows the sea-ice record superimposed on the inverted time series of the second EOF of winter (November to March) SST (see Figure 1). Both the sea-ice and the SST time series have been detrended by subtracting least-squares parabolas based on the period 1950 to 1988, and smoothed with a 3-point binomial filter. It may be seen that the maxima in sea-ice concentration precede the minima in SST by one to two years. The correlation between the two time series is -0.26 with no lag, -0.62 when sea ice leads SST by 1 year, -0.76 when sea ice leads by 2 years, and -0.62 when sea ice leads by 3 years. The strong lag correlations indicate that, on the decadal time scale, winters of heavy sea ice in the Labrador Sea precede winters of colder-than-normal SSTs east of Newfoundland. It is plausible that the sea-ice anomalies in the Labrador Sea are advected southeastward, resulting in colder-than-normal SSTs east of Newfoundland the following (or second) year. Thus, the quasi-decadal cycle in SSTs east of Newfoundland may result in part from low-frequency Arctic sea-ice variations. Mysak and Manak (1989) and Mysak (1991) have also discussed the possible link between sea ice and SST anomalies in the northern North Atlantic, and Mysak et al. (1990) have postulated the existence of an interdecadal Arctic climate cycle involving
runoff, sea ice, and SSTs. Further analysis is needed to elucidate the role of Arctic sea ice in decadal climate variability over the North Atlantic.
Another dominant mode of variability of the wintertime