GREENHOUSE GASES EFFECTS 433
not surprisingly, show a greater spread, from 5 to 15% with
an average of 10%.
These predictions were not much affected by doubling
the horizontal resolution (having the grid spacing). However,
they were much more sensitive to the formulation of physi-
cal processes, in particular the representation of clouds and
their interactions with solar and terrestrial radiation. Model
simulations in which the cloud water was computed from
the model variables and their radiative properties (emissiv-
ity, absorptivity and reflectivity) were allowed to vary with
the liquid water and ice content produced significantly dif-
ferent results as summarized in Table 1.
The UKMO model, using three progressively more sophis-
ticated and realistic cloud/radiation schemes, has progressively
reduced the predicted global warming from 5.2K to 1.9K and
the corresponding precipitation increases from 15% to 3%. It
is important to identify and understand the underlying physical
reasons for these results which, if confirmed, are likely to have
an important influence on the whole GHW debate.
In the first version of the model, in which cloud cover
was related empirically only to relative humidity and the
radiative properties were fixed during the whole simula-
tion, enhanced CO 2 produced unrealistic decreases in high-,
medium- and low-level clouds, except at very high latitudes
and, consequently, an exaggerated warming of the atmo-
sphere. Decrease in cloud amount seems inconsistent with
the predicted increase in precipitation and suggests that the
empirically derived cloud cover was incompatible with the
internal dynamics of the model. In the most sophisticated
treatment, the cloud water is computed from the dynamical
and physical equations; it is transformed progressively from
liquid water to ice as the temperature falls below −15C;
rapidly growing ice crystals are allowed to fall out of the
cloud; and the radiative properties are varied as a function
of the cloud water path and the solar angle for the incoming
solar radiation and as a function of the water/ice path for
terrestrial long-wave radiation. In this case, enhanced CO 2
leads to a marked increase in the extent and optical depth of
call clouds, and especially of low clouds in middle and high
latitudes, which reflect more of the solar radiation to space
and therefore reduce the GHW of the atmosphere to only
1.9K. The small 3% increase in precipitation is consistent
with a 2–3% increase in low cloud cover and a 2% increase
in medium-level cloud in the Northern Hemisphere. A more
detailed account is given by Senior and Mitchell (1993).
Transient Experiments in Which CO 2
Increases at 1% p.a.
The fact that we now have fully three-dimensional models of the
global oceans coupled interactively to the atmosphere, land-
surface and sea-ice components of the climate model, enabling
FIGURE 4 Prediction of the UKMO coupled atmosphere—deep ocean model of global warming caused by increasing
the concentration of atmospheric carbon dioxide by 1% p.a. compound after 75 years.
COUPLED MODEL
10 YEAR ANNUAL MEAN (YEARS 66 TO 75)
SURFACE AIR TEMPERATURE
<–2 –2 to 0 C 0 to 1 C 1 to 2 C 2 to 4 C >4 C
C007_002_r03.indd 433C007_002_r03.indd 433 11/18/2005 10:28:00 AM11/18/2005 10:28:00 AM