386
warm periods (1925-39 and 50-64) during this century. He found that the
difference between warm and cold years consists of positive SST anomalies
along the east coast of Greenland, the entrance of the Labrador Sea and
off-shore of Newfoundland/Nova Scotia and negative SST anomalies along
the northern part of the US east coast (Fig. 3). This is associated with a
basin-wide cyclonic sea level pressure anomaly very similar to that reported
by Deser and Blackmon (1993). The SST anomaly pattern is present in
both the warm and the cold season (Fig. 3a, b), whereas the pressure
anomaly is primarily a winter time phenomenon. While in the southern
region of this SST anomaly the atmospheric circulation anomalies tend to
attenuate the SST anomalies, they are maintained by them in the northern region. Hydrographic changes in the shallow and intermediate North
Atlantic reported by Levitus (1989a), Levitus (1989b) are consistent with
these findings. A similar structure (cooling at the Labrador Sea entrance
and east coast of Greenland and warming is again found in the subsurface
temperature (125 m) trends from 1966 to 1990 (Levitus et al. 1994)).
It is very important to address the question whether the North Atlantic
is a region of enhanced variability as suggested by some proxy records or
whether this is solely a result of the sparsity of data elsewhere. Schlesinger
and Ramankutty (1994) analyze global data sets of surface air temperature
covering the period 1858-1992 in 11 different geographical regions. Using
singular spectrum analysis, they found a 50-80 year cycle in this variable
in various sub-regions of the globe with the largest amplitudes in the North
Atlantic and North American regions. This is an important hint towards
an underlying mechanism. Similar decadal-to-century cycles at 24 and 100
years have been found in the Central England temperature time series,
the longest instrumental record available, covering 318 years (Stocker and
Mysak 1992). They are statistically significant at the 99% level.
2.2 Proxy Data
A review of long-term cyclic fluctuations on the century time scale is given
in Stocker and Mysak (1992). Cycles of 50 years and longer are abundant
in high-resolution proxy records (Fig. 4), but a clear and unequivocal time
scale is missing. Stuiver (1980) tested the hypothesis whether such cycles
found in proxy data could be due to solar variability for which ,6. 14 C from
tree rings was used as a proxy. A statistically significant relationship could
not be established.
warm periods (1925-39 and 50-64) during this century. He found that the
difference between warm and cold years consists of positive SST anomalies
along the east coast of Greenland, the entrance of the Labrador Sea and
off-shore of Newfoundland/Nova Scotia and negative SST anomalies along
the northern part of the US east coast (Fig. 3). This is associated with a
basin-wide cyclonic sea level pressure anomaly very similar to that reported
by Deser and Blackmon (1993). The SST anomaly pattern is present in
both the warm and the cold season (Fig. 3a, b), whereas the pressure
anomaly is primarily a winter time phenomenon. While in the southern
region of this SST anomaly the atmospheric circulation anomalies tend to
attenuate the SST anomalies, they are maintained by them in the northern region. Hydrographic changes in the shallow and intermediate North
Atlantic reported by Levitus (1989a), Levitus (1989b) are consistent with
these findings. A similar structure (cooling at the Labrador Sea entrance
and east coast of Greenland and warming is again found in the subsurface
temperature (125 m) trends from 1966 to 1990 (Levitus et al. 1994)).
It is very important to address the question whether the North Atlantic
is a region of enhanced variability as suggested by some proxy records or
whether this is solely a result of the sparsity of data elsewhere. Schlesinger
and Ramankutty (1994) analyze global data sets of surface air temperature
covering the period 1858-1992 in 11 different geographical regions. Using
singular spectrum analysis, they found a 50-80 year cycle in this variable
in various sub-regions of the globe with the largest amplitudes in the North
Atlantic and North American regions. This is an important hint towards
an underlying mechanism. Similar decadal-to-century cycles at 24 and 100
years have been found in the Central England temperature time series,
the longest instrumental record available, covering 318 years (Stocker and
Mysak 1992). They are statistically significant at the 99% level.
2.2 Proxy Data
A review of long-term cyclic fluctuations on the century time scale is given
in Stocker and Mysak (1992). Cycles of 50 years and longer are abundant
in high-resolution proxy records (Fig. 4), but a clear and unequivocal time
scale is missing. Stuiver (1980) tested the hypothesis whether such cycles
found in proxy data could be due to solar variability for which ,6. 14 C from
tree rings was used as a proxy. A statistically significant relationship could
not be established.