Global warming, models and language - Climate Change: The Facts - page 25

basedontheagreement ofalmostallscientists withthefourtrivialyes-nostatements atthe
beginning of this essay.
It is relatively simple to show how such considerations play out in practice using what
arecalledenergybalancemodels.IncontrasttoGeneralCirculationModels(GCMs)where
climatesensitivityisinternallygenerated,thesemodelsassumeasensitivityandinvestigate
the response of the surface temperature where heat capacity is that of a relatively simple
ocean model. When the chosen sensitivity is that of a GCM, the resulting evolution of
temperaturefollowsthatoftheGCM.Asaresult,thesesimplemodelsareusedextensively
by the IPCC for scenario building. The specific model we will use is that described in
Richard Lindzen and Constantine Giannitsis' 1998 paper, 'On the climatic implications of
volcanic cooling'. 2 Before going into details, it is worth pointing out that these models
describe an important feature of climate: namely, that high sensitivities are associated with
longresponsetimes.Radiativeforcingisafluxofenergy(i.e.anenergyflowperunitarea),
and sensitivity is a ratio of temperature change to this flux. High sensitivity means that a
small flux eventually produces a large temperature change, but, because the flux is small,
the change will take a long time.
Our exercise begins with two sources of forcing common to almost all models. The
first is the radiative forcing due to anthropogenic greenhouse gases (CO 2 , methane, nitrous
oxide, etc.). According to the IPCC, this has grown from about zero in 1850 to about 3
wattspersquaremeter(Wm -2 )now.ThishasresultedfromtheexponentialgrowthofCO 2 .
Note that we are already at about 80 per cent of what one expects from a doubling of
CO 2 alone. The other is the forcing due to major volcanoes. Figure 1 shows the volcanic
forcing estimated by Makiko Sato. 3 The values are comparable to greenhouse forcing, but
thevolcanoesoccurintwoclusterswithaninterveningquiescentperiod.Suchclusteringis
actually characteristic of random processes.
Figure 2a shows the response of the model for various values of climate sensitivity.
The observed warming has been about 0.75C. This is exceeded by almost any sensitivity
in excess of 1C. However, the calculated responses for models with high sensitivity are
smaller than the equilibrium responses because of the long response times associated with
high sensitivity.