141
0.6 ,---~----.---~--~----.---~---.
6" 0.4
e...
0.2
~
:::I
-.... Cll
Q) -0.2
0..
E
~ -0.4
(a)
•
-0.6+----.----,----.----.---~----~--~
1955 1960 1965 1970 1975 1980 1985 1990
Year
O. 6 ,-----.---...--___,_--r--~-...,.._-_,
6" 0.4
e...
c:
o
1ii
'> Q)
"0
~
:::I
0.2
"§
Q) -0.2
0..
E
~ -0.4
(b)
-0.6+----.-----,-----.---~---...------,---~
1955 1960 1965 1970 1975 1980 1985 1990
Year
Figure 28: Observed temperature anomalies in the troposphere and lower stratosphere
from 1958- 1989, based on Angell (1988). a) annual global values for 850-300hPa. Dots
are values from Spencer and Christy (1990). b) 300-100hPa. From IPCC (1992).
This appears to be consistent with the predictability of the THe index in
Griffies and Bryan's (1996) study, shown in Fig 25.
It is worth noting that a model such as (5.1) could never be considered
a representation of ENSO. For example, the tropical atmosphere is not, by
itself, chaotic, and internal atmospheric dynamics alone cannot account for
the Southern Oscillation (Dix and Hunt, 1995). Moreover, it is well known
that ocean dynamics are essential for the ENSO mechanism, and therefore
cannot be represented by a simple storage effect.