I
!
1910
40
1930
us
1950
1970
1990
Figure 3: As in the lower two panels of Fig. 2 but the data have been smoothed with
a 5-year running mean filter. The spacing between the curves is arbitrary. The interval
between tick marks on the vertical axis is 0.25 K.
.. . . 0 • •
• • 0
.0
. . . . .
• , .
o!
~ •• : ·.0 .. 0. : eo 0. :'!
. ::- : ... : · . . . :.t~.:·: .. ;·~ . . :o.!-:.:::r .. ::
•
eo ... ,
~.":: • d\. -: . . . . . . . . . :0.: -'t ••• -oJ.: . -0 •• ~.: .-.:':, ••• t': :.~ • ~." •• t.-'':':~. ! ,. ....
.................... :.... .:.
••
.r..
: ........ ~ . .: ..... 'J.: ,.:.::- .:.~; .... =:-::. -:.: .
: : ••• "
• • : : • • • : . . . . . . . . . . . .
~ . . . . •
: • • 0
•
.0.:. eo . . . . :
.:.00 : -0· ..
0.. ' •. -.0. 0
..
~
g
1
0
f-<
-1
' ..
MAY-OCT
NOV-APR
-2
1910
1930
1950
1970
1990
1910
1930
1950
1970
1990
Figure 4: Northern Hemispheric-mean surface air temperature anomalies based on gridded
station data, obtained from the Hadley Centre for Climate Prediction and Research. The
data are partitioned into Northern Hemisphere warm and cold seasons, as indicated. From
Wallace et a1. (1995a)
warm and cold seasons have exhibited similar patterns of interdecadal to
century scale variability. The same is true of surface air temperature over
the Southern Hemisphere continents (Wallace, 1995).
From the foregoing, it is apparent that the structure of climate variability on the interannual to interdecadal time scale is quite complex, but
the evidence presented here does not contradict the prevailing view that it
may be much simpler on the time scale of centuries or longer. In order to
understand the nature of the month-to-month variability in local, regional
and hemispheric mean surface air temperature and to explain why it is so
much larger during the cold season, it will be necessary to learn something
about the horizontal structure of the geopotential height and temperature
patterns in the troposphere. Since the processes that generate tropical
!
1910
40
1930
us
1950
1970
1990
Figure 3: As in the lower two panels of Fig. 2 but the data have been smoothed with
a 5-year running mean filter. The spacing between the curves is arbitrary. The interval
between tick marks on the vertical axis is 0.25 K.
.. . . 0 • •
• • 0
.0
. . . . .
• , .
o!
~ •• : ·.0 .. 0. : eo 0. :'!
. ::- : ... : · . . . :.t~.:·: .. ;·~ . . :o.!-:.:::r .. ::
•
eo ... ,
~.":: • d\. -: . . . . . . . . . :0.: -'t ••• -oJ.: . -0 •• ~.: .-.:':, ••• t': :.~ • ~." •• t.-'':':~. ! ,. ....
.................... :.... .:.
••
.r..
: ........ ~ . .: ..... 'J.: ,.:.::- .:.~; .... =:-::. -:.: .
: : ••• "
• • : : • • • : . . . . . . . . . . . .
~ . . . . •
: • • 0
•
.0.:. eo . . . . :
.:.00 : -0· ..
0.. ' •. -.0. 0
..
~
g
1
0
f-<
-1
' ..
MAY-OCT
NOV-APR
-2
1910
1930
1950
1970
1990
1910
1930
1950
1970
1990
Figure 4: Northern Hemispheric-mean surface air temperature anomalies based on gridded
station data, obtained from the Hadley Centre for Climate Prediction and Research. The
data are partitioned into Northern Hemisphere warm and cold seasons, as indicated. From
Wallace et a1. (1995a)
warm and cold seasons have exhibited similar patterns of interdecadal to
century scale variability. The same is true of surface air temperature over
the Southern Hemisphere continents (Wallace, 1995).
From the foregoing, it is apparent that the structure of climate variability on the interannual to interdecadal time scale is quite complex, but
the evidence presented here does not contradict the prevailing view that it
may be much simpler on the time scale of centuries or longer. In order to
understand the nature of the month-to-month variability in local, regional
and hemispheric mean surface air temperature and to explain why it is so
much larger during the cold season, it will be necessary to learn something
about the horizontal structure of the geopotential height and temperature
patterns in the troposphere. Since the processes that generate tropical