Geoscience Reference
In-Depth Information
General circulation models
APPLICATIONS
The rapid development of computing power has made it possible to model the physical processes which operate in
the atmosphere and the oceans and to simulate Earth's atmospheric circulation realistically. Such dynamic models
are called general circulation models (GCMs).
A GCM uses mathematics and the laws of physics to describe the operation of the atmosphere. In summary, the
model is started off with a known climatology, usually resembling that of the present Earth. The data are provided
for a grid network with horizontal separation of several hundred kilometres (usually 3
longitude) and
information for several heights into the atmosphere for the vertical resolution; more information is obtained about
the lower levels of the atmosphere than about the higher levels. The solar input and radiational output are readily
known and the main problem is to model the relationship between the surface, the atmosphere and the oceans. To
do so, a number of assumptions about and simplifications of the interactions have to be incorporated into the model
(
Figure 6.16
).
An example is the role of clouds, which was discussed in
Chapter 4.
A major complication is how to
link the rapid atmospheric movement with the much slower circulation of the oceans and even slower responses in
ice sheets. Early models used fixed sea surface temperatures based on present-day values which were allowed to
vary seasonally or incorporated the meridional energy transport of the oceans. The latest models can allow for vertical
and horizontal exchanges in the oceans to give more realistic results. Close interaction between the two subsystems
of air and ocean is impossible because the ocean layer needs a much longer time to reach equilibrium from any given
change. That is why sea surface temperature anomalies are more persistent than those of the atmosphere.
latitude by 3
General circulation models simulate the behaviour of the real atmosphere and reproduce the main circulation features
outlined in this chapter. Even individual weather systems are generated by the computer model. The models can
either be used for short-period weather prediction extending to about ten days ahead, or they can be modified for
climate prediction. In that case the model is run to simulate several decades, to ensure that it reproduces the real
atmosphere adequately. Once it is in equilibrium, a variable may be changed. We could alter the concentration of
carbon dioxide or the nature of the ground surface to simulate Amazonian deforestation. The model is then run
repeatedly with increasing levels of carbon dioxide or reduced areas of forest to see what the effect on the circulation
would be. A novel use of GCMs is to attempt to reproduce former circulations. With improved observational techniques
worldwide climatic records are being obtained from soils, lake and ocean sediments and ice strata (see
Chapter 23).
This wealth of knowledge can be used to infer the nature of the atmosphere and ground surface conditions in the
past. It is now possible to allow for changes in Earth's orbit around the sun, to simulate the effect of increased areas
of ice at the surface during the last Ice Age, and even to change the location of the continents to determine what
their impact might be.
Although GCMs have a number of limitations they are at present the best way of estimating possible climate change.
Developments are taking place in two ways. First, improvements in computer power will allow us to incorporate
more information and make calculations even more quickly. In that way the horizontal and vertical resolution of a
model can be increased so that the initial state of the systems can be portrayed more precisely. Alternatively, more
detailed models can be 'nested' within the larger model to examine finer detail. For example, the highest-resolution
UK Regional Climate Model now uses grid squares just 10 km square, compared with 150 km squares for the global
model. Second, the modelling of the interaction of air, land, ice and water needs to be improved, perhaps with the
incorporation of chemical interactions such as the changes in stratospheric ozone.
Numerical modelling of the global climate system has led to a better understanding of how it works. As human
activities may be altering the climate, it is of vital importance to be reasonably certain what the implications are. It
will require major resources in computers, observational programmes and scientific research - but at least we are
much further along the line of progress than thirty years ago.
