51
flow (Simmons et al. 1983) and the patterns most strongly excited by
random heat and vorticity sources in a linear multi-level GCM (Branstator 1990). Furthermore, they strongly resemble the dominant modes of
low-frequency variability in a variety of GCM's, run with and without
time varying bottom boundary conditions. They prevail because they are
particularly effective in extracting kinetic energy from the zonally varying
climatological mean wintertime flow. The eastward, down-gradient flux of
zonal momentum in the jet exit regions is instrumental in the energy conversion process (Simmons et al. 1983, Wallace and Lau 1985, Nakamura et
al. 1987). Hence, there can be little doubt that these modes and the related
patterns identified by cluster analysis reflect natural, dynamical variability
of the atmospheric component of the climate system. They exist independently of any time-varying thermodynamic forcing of the atmosphere such
as greenhouse warming or variations in sea surface temperature or land
surface properties.
In comparison to the wintertime field, the summertime 500-mb height
field is relatively quiescent and featureless. Throughout most of the hemisphere, the root-mean-squared variance of summertime mean 500-mb height
is only on the order of half as large as that of wintertime mean 500-mb
height (Wallace et al. 1993, Fig. 1), and the leading EOF tends to be
of the same polarity throughout nearly the entire hemisphere, in marked
contrast to the more wavelike patterns in Figs. 7-10. Yet, despite its unaesthetic appearance when rendered as a contour map (ibid., Fig. 2), this
EOF explains almost as much of the variance of the 500-mb height field
as its wintertime counterpart. Such bland, monopolar patterns are even
more prominent among the EOFs of the hemispheric 1000-500-mb thickness field, which is representative of vertically averaged lower tropospheric
temperature; i.e., they account for larger fractions of the total variance
than their counterparts in the 500-mb height field (ibid., Table 2). The
time series of the expansion coefficients of such EOFs are highly correlated
with the time series of the hemispheric-mean of the field itself. The strong
seasonality of these structures has implications for the interpretation of
hemispheric-mean temperature trends, as will be discussed in the next two
sections.
flow (Simmons et al. 1983) and the patterns most strongly excited by
random heat and vorticity sources in a linear multi-level GCM (Branstator 1990). Furthermore, they strongly resemble the dominant modes of
low-frequency variability in a variety of GCM's, run with and without
time varying bottom boundary conditions. They prevail because they are
particularly effective in extracting kinetic energy from the zonally varying
climatological mean wintertime flow. The eastward, down-gradient flux of
zonal momentum in the jet exit regions is instrumental in the energy conversion process (Simmons et al. 1983, Wallace and Lau 1985, Nakamura et
al. 1987). Hence, there can be little doubt that these modes and the related
patterns identified by cluster analysis reflect natural, dynamical variability
of the atmospheric component of the climate system. They exist independently of any time-varying thermodynamic forcing of the atmosphere such
as greenhouse warming or variations in sea surface temperature or land
surface properties.
In comparison to the wintertime field, the summertime 500-mb height
field is relatively quiescent and featureless. Throughout most of the hemisphere, the root-mean-squared variance of summertime mean 500-mb height
is only on the order of half as large as that of wintertime mean 500-mb
height (Wallace et al. 1993, Fig. 1), and the leading EOF tends to be
of the same polarity throughout nearly the entire hemisphere, in marked
contrast to the more wavelike patterns in Figs. 7-10. Yet, despite its unaesthetic appearance when rendered as a contour map (ibid., Fig. 2), this
EOF explains almost as much of the variance of the 500-mb height field
as its wintertime counterpart. Such bland, monopolar patterns are even
more prominent among the EOFs of the hemispheric 1000-500-mb thickness field, which is representative of vertically averaged lower tropospheric
temperature; i.e., they account for larger fractions of the total variance
than their counterparts in the 500-mb height field (ibid., Table 2). The
time series of the expansion coefficients of such EOFs are highly correlated
with the time series of the hemispheric-mean of the field itself. The strong
seasonality of these structures has implications for the interpretation of
hemispheric-mean temperature trends, as will be discussed in the next two
sections.