Geoscience Reference
In-Depth Information
Figure 13.15
Plot of mean annual temperatures for the northern hemisphere for the period 1880 to 1980, to which have been fitted
a linear trend and a sinusoidal oscillation. The latter have been extended to attempt a prediction to the year 2100.
Figure 13.16
A grey box model of
global mean warming from 1850 to
1990 using greenhouse gas forcing
alone and greenhouse gas plus anthro-
pogenic sulphate aerosol forcing. It is
clear that the latter provides much the
better fit to the observed temperature
record.
Source
: Houghton
et al
. (1996).
Reproduced by permission of the IPCC
and Cambridge University Press.
model is capable of being projected meaningfully
into the future. Figure 13.16 shows one such
attempt involving the variation of atmospheric CO
2
concentration and sulphate aerosols in the atmos-
phere. The difficulties of this type of approach to
prediction are, again, quite obvious and fall into two
categories:
•
The accurate prediction of future changes in CO
2
concentration is difficult and that of sulphate and
other aerosols is even more uncertain.
It is clear that the complexity of the global climate
system is such as to rule out the use of statistically or
empirically based models, with their total reliance
on past data, in favour of numerical ones.
3
White box modelling
is based on detailed under-
standing of the structure and operation of the earth-
atmosphere-ocean system, such that its possible
future states can be simulated by applying assumed
forcing mechanisms, particularly anthropogenic
ones. Numerical model building involves bringing
together locational, temporal and attribute infor-
mation into a database that allows climatic processes
•
The fitting of a composite curve to the past record
is uncertain. Have all the important variables been
included? Have their actual and relative effects
on hemispheric temperature been accurately
determined? Is it legitimate to assume that the
effect of each variable operates independently of
that of the others to allow the cumulative effect to
be calculated?
