273
and Central America. The anomalies associated with the PNA pattern are
statistically highly significant, with explained variances relative to the lowpass filtered height anomalies of about 90% in the center of the Aleutian
anomaly and about 60% in the centers further downstream.
3.2 Coupled general circulation model simulation
The corresponding results for the coupled general circulation model are
shown in Fig. 2. We have chosen a model period that begins in year 43 and
extends to year 83 for the correlation analysis. This time period shows a
very similar time behaviour in North Pacific SST as the observations during the period 1950-1990 (compare Fig. 2a with Fig. 1a). Overall, the
coupled model results are consistent with the observations. The associated SST correlation pattern as derived from the CGCM (Fig. 2b) shows
the main positive anomaly in the North Pacific which is surrounded by
negative anomalies. The symmetry about the equator is also found in the
coupled model simulation. The negative equatorial anomaly, however, is
less pronounced than in the observations.
We use the 500-hPa field to describe the changes in the large-scale atmospheric circulation in the coupled simulation. Since the changes in the
height field are equivalent barotropic, the model results (Fig. 2c) can be
compared directly to the observations (Fig. 1c), although they are presented at a different vertical level. The atmospheric response in the coupled model is also characterized by the reverse of the PNA pattern, but
the changes are systematically too zonal relative to the observations. The
fundamental spatial phase relationship between the North Pacific SST and
height anomaly fields, however, are very similar: anomalously warm SSTs
are accompanied by anomalously high pressure over the North Pacific and
thus a weakened Aleutian low.
Thus, we believe that our CGCM simulates reasonably well the decadal
variations observed in the North Pacific, so that we can use it to investigate the dynamics of the decadal variability in more detail. This is not
possible with the available observations which are too sparse and not homogeneous. To elucidate the mechanism producing the decadal variations
in SST, we investigated the characteristic evolution of upper ocean heat
content anomalies, as defined by the vertically averaged temperatures over
the upper 500 meters of the water column via a complex empirical orthogonal function (CEOF) analysis (Barnett (1983)). Before the CEOF
analysis, the heat content data were smoothed with a low-pass filter that
retained variability at time scales longer than 3 years (further details can
and Central America. The anomalies associated with the PNA pattern are
statistically highly significant, with explained variances relative to the lowpass filtered height anomalies of about 90% in the center of the Aleutian
anomaly and about 60% in the centers further downstream.
3.2 Coupled general circulation model simulation
The corresponding results for the coupled general circulation model are
shown in Fig. 2. We have chosen a model period that begins in year 43 and
extends to year 83 for the correlation analysis. This time period shows a
very similar time behaviour in North Pacific SST as the observations during the period 1950-1990 (compare Fig. 2a with Fig. 1a). Overall, the
coupled model results are consistent with the observations. The associated SST correlation pattern as derived from the CGCM (Fig. 2b) shows
the main positive anomaly in the North Pacific which is surrounded by
negative anomalies. The symmetry about the equator is also found in the
coupled model simulation. The negative equatorial anomaly, however, is
less pronounced than in the observations.
We use the 500-hPa field to describe the changes in the large-scale atmospheric circulation in the coupled simulation. Since the changes in the
height field are equivalent barotropic, the model results (Fig. 2c) can be
compared directly to the observations (Fig. 1c), although they are presented at a different vertical level. The atmospheric response in the coupled model is also characterized by the reverse of the PNA pattern, but
the changes are systematically too zonal relative to the observations. The
fundamental spatial phase relationship between the North Pacific SST and
height anomaly fields, however, are very similar: anomalously warm SSTs
are accompanied by anomalously high pressure over the North Pacific and
thus a weakened Aleutian low.
Thus, we believe that our CGCM simulates reasonably well the decadal
variations observed in the North Pacific, so that we can use it to investigate the dynamics of the decadal variability in more detail. This is not
possible with the available observations which are too sparse and not homogeneous. To elucidate the mechanism producing the decadal variations
in SST, we investigated the characteristic evolution of upper ocean heat
content anomalies, as defined by the vertically averaged temperatures over
the upper 500 meters of the water column via a complex empirical orthogonal function (CEOF) analysis (Barnett (1983)). Before the CEOF
analysis, the heat content data were smoothed with a low-pass filter that
retained variability at time scales longer than 3 years (further details can