147
in hemispheric mean temperature, the large-scale (nonlinear) dynamics of
the atmosphere appear to lead us to focus attention onto the possible
impact of enhanced CO2 on key physical processes in specific localised sensitive regions of the globe; the very antithesis of the global greenhouse
effect.
It can be noted that in fact SSTs have increased over the west Pacific
and Indian Ocean over the last few decades. Fig 29a (from Latif et al,
1996) shows the linear trend in SST (deg Cjyr) over the period 1949-1991,
Fig 29b shows the percentage of variance that this linear trend explains.
It can be seen that the trend over the West Pacific and the Indian Ocean
explains a relatively large part of the observed variance in SST in those
areas.
Now what has this discussion to do with the predictability of climate
change? According to these ideas, if we are to be able to simulate accurately the global atmospheric response to enhanced CO2, it may be necessary to simulate very accurately how CO2 influences the atmosphere in
very specific regions of the globe (specifically over the warm pool area),
and perhaps less accurately elsewhere. At present, uncertainties in basic
radiative flux parametrisations in the warm pool area certainly exceed the
basic 4W jm 2 associated with radiative effect of doubled CO2.
Despite these remarks, the predictability of climate change may be
greater in the northern summer. In summer, the role of regime dynamics
may be weakest, and the radiative greenhouse mechanism may be most
significant. Moreover, according to the results of Wallace et al (1996) the
amorphous component of hemispheric temperature change has its largest
amplitude in summer.
Let us conclude this section by asking why the IPCC GCM simulations
have not replicated the Angell observations? It is possible that the largescale dynamics of many GCMs is still inadequate, and that the simulated
warming in the 300-100hPa region is a manifestation of this inadequacy.
For example, stationary-wave amplitudes in many GCMs are still poor,
although recent diagnoses have suggested that weather regime structure
does exist in the latest generation of atmospheric climate models (Haines
and Hannachi, 1995). My guess is that as models improve, particularly
in their representation of low-frequency atmospheric variability, the vertical structure of the response of GCMs to observed CO2 will correspond
more and more closely to the observed vertical structure of warming. (Of
course I am aware that there are other complicating factors in discussing
in hemispheric mean temperature, the large-scale (nonlinear) dynamics of
the atmosphere appear to lead us to focus attention onto the possible
impact of enhanced CO2 on key physical processes in specific localised sensitive regions of the globe; the very antithesis of the global greenhouse
effect.
It can be noted that in fact SSTs have increased over the west Pacific
and Indian Ocean over the last few decades. Fig 29a (from Latif et al,
1996) shows the linear trend in SST (deg Cjyr) over the period 1949-1991,
Fig 29b shows the percentage of variance that this linear trend explains.
It can be seen that the trend over the West Pacific and the Indian Ocean
explains a relatively large part of the observed variance in SST in those
areas.
Now what has this discussion to do with the predictability of climate
change? According to these ideas, if we are to be able to simulate accurately the global atmospheric response to enhanced CO2, it may be necessary to simulate very accurately how CO2 influences the atmosphere in
very specific regions of the globe (specifically over the warm pool area),
and perhaps less accurately elsewhere. At present, uncertainties in basic
radiative flux parametrisations in the warm pool area certainly exceed the
basic 4W jm 2 associated with radiative effect of doubled CO2.
Despite these remarks, the predictability of climate change may be
greater in the northern summer. In summer, the role of regime dynamics
may be weakest, and the radiative greenhouse mechanism may be most
significant. Moreover, according to the results of Wallace et al (1996) the
amorphous component of hemispheric temperature change has its largest
amplitude in summer.
Let us conclude this section by asking why the IPCC GCM simulations
have not replicated the Angell observations? It is possible that the largescale dynamics of many GCMs is still inadequate, and that the simulated
warming in the 300-100hPa region is a manifestation of this inadequacy.
For example, stationary-wave amplitudes in many GCMs are still poor,
although recent diagnoses have suggested that weather regime structure
does exist in the latest generation of atmospheric climate models (Haines
and Hannachi, 1995). My guess is that as models improve, particularly
in their representation of low-frequency atmospheric variability, the vertical structure of the response of GCMs to observed CO2 will correspond
more and more closely to the observed vertical structure of warming. (Of
course I am aware that there are other complicating factors in discussing