44
Figure 7: One-point correlation maps based on the grid-point (45 N, 165 W) which
corresponds to the primary centre of action of the PNA pattern. Based on (left to right)
unfiltered, 10-day lowpass filtered, and monthly-mean NMC operational analyses for the
winter months (December through February) of the years 1946-93. Contour interval 0.2.
Figure 8: The first (left) and second (right) EOF's of wintertime-mean (December through
February) 500-mb height, computed from the temporal covariance matrix, based on 10day lowpass filtered NMC operational analyses for the period 1946-89. Contour interval
20m.
day lowpass filtered 500-mb height field, shown in Fig. 8. Since EOF's, by
their very definition, are structured so as to explain as much as possible
of the variance, it is to be expected that the primary "center of action" of
the leading EOF should coincide with the region of strongest variability.
Are the correlation patterns in Fig. 7 typical of correlation patterns for
gridpoints throughout the hemisphere, or are they particularly strong ones?
The most straightforward way to answer this question is to systematically
examine the complete set of one-point correlation maps analogous to the
ones in Fig. 7, but for all grid points. Such a survey has, in fact been
Figure 7: One-point correlation maps based on the grid-point (45 N, 165 W) which
corresponds to the primary centre of action of the PNA pattern. Based on (left to right)
unfiltered, 10-day lowpass filtered, and monthly-mean NMC operational analyses for the
winter months (December through February) of the years 1946-93. Contour interval 0.2.
Figure 8: The first (left) and second (right) EOF's of wintertime-mean (December through
February) 500-mb height, computed from the temporal covariance matrix, based on 10day lowpass filtered NMC operational analyses for the period 1946-89. Contour interval
20m.
day lowpass filtered 500-mb height field, shown in Fig. 8. Since EOF's, by
their very definition, are structured so as to explain as much as possible
of the variance, it is to be expected that the primary "center of action" of
the leading EOF should coincide with the region of strongest variability.
Are the correlation patterns in Fig. 7 typical of correlation patterns for
gridpoints throughout the hemisphere, or are they particularly strong ones?
The most straightforward way to answer this question is to systematically
examine the complete set of one-point correlation maps analogous to the
ones in Fig. 7, but for all grid points. Such a survey has, in fact been