41
and extratropical variability are somewhat different, these regions will be
considered in separate sections. The discussion in these sections will also
provide some insights into the nature of regional climate variability on the
interdecadal time scale and the extent to which it is coupled to variations
in hemispheric- or global-mean temperature.
4 Horizontal structure of extratropical climate variability
The structure of the low-frequency (lO-day period and longer) variability of
the extratropical circulation is strongly seasonally dependent. Therefore
it will be convenient to consider the winter and summer seasons separately. We will begin with winter, whose climatological mean circulation
is stronger because of the stronger meridional gradient of diabatic heating
that drives it.
The Northern Hemisphere climatological mean wintertime circulation,
as reflected in the 500-mb height field, shown in Fig. 5, exhibits strong
zonal asymmetries forced primarily by the continent-ocean heating contrasts and the presence of the Rockies and the Himalayas. These asymmetries give rise to a strong longitudinal dependence in the amplitude and
structure of the observed low-frequency variability. Figure 6 shows the geographical distribution of the temporal variance of the 500-mb height field
based on unfiltered daily data, lO-day low-pass filtered daily data, monthly
means and seasonal (December through February) means. All four patterns
exhibit pronounced variance maxima over the northern oceans. Variance
decreases with the strength of the temporal smoothing that is applied to
the data. The four patterns are similar, but upon careful inspection it is
evident that the patterns based on monthly and seasonal means exhibit
relatively stronger longitudinal contrasts, in a relative sense, than the less
heavily smoothed patterns. Although it doesn't show up clearly in the figure, the variance maximum over the North Pacific stands out particularly
strongly (in a relative sense) in the seasonal-mean data.
Figure 7 shows a series of one-point correlation maps for the gridpoint
(45°N, l65°W) which is centered near the region of largest variance in
the monthly and seasonal panels of the previous figure. These maps are
constructed by correlating the 500-mb height time series at the 'reference
gridpoint' with the corresponding time series at all gridpoints. The one in
and extratropical variability are somewhat different, these regions will be
considered in separate sections. The discussion in these sections will also
provide some insights into the nature of regional climate variability on the
interdecadal time scale and the extent to which it is coupled to variations
in hemispheric- or global-mean temperature.
4 Horizontal structure of extratropical climate variability
The structure of the low-frequency (lO-day period and longer) variability of
the extratropical circulation is strongly seasonally dependent. Therefore
it will be convenient to consider the winter and summer seasons separately. We will begin with winter, whose climatological mean circulation
is stronger because of the stronger meridional gradient of diabatic heating
that drives it.
The Northern Hemisphere climatological mean wintertime circulation,
as reflected in the 500-mb height field, shown in Fig. 5, exhibits strong
zonal asymmetries forced primarily by the continent-ocean heating contrasts and the presence of the Rockies and the Himalayas. These asymmetries give rise to a strong longitudinal dependence in the amplitude and
structure of the observed low-frequency variability. Figure 6 shows the geographical distribution of the temporal variance of the 500-mb height field
based on unfiltered daily data, lO-day low-pass filtered daily data, monthly
means and seasonal (December through February) means. All four patterns
exhibit pronounced variance maxima over the northern oceans. Variance
decreases with the strength of the temporal smoothing that is applied to
the data. The four patterns are similar, but upon careful inspection it is
evident that the patterns based on monthly and seasonal means exhibit
relatively stronger longitudinal contrasts, in a relative sense, than the less
heavily smoothed patterns. Although it doesn't show up clearly in the figure, the variance maximum over the North Pacific stands out particularly
strongly (in a relative sense) in the seasonal-mean data.
Figure 7 shows a series of one-point correlation maps for the gridpoint
(45°N, l65°W) which is centered near the region of largest variance in
the monthly and seasonal panels of the previous figure. These maps are
constructed by correlating the 500-mb height time series at the 'reference
gridpoint' with the corresponding time series at all gridpoints. The one in