WMOD00435

• Title: Limits on Global Warming
• Date: 5/9/1991
• Weather Modification - Doc Type: Report

analogue for a 2 to 2.5 oC warming; and the Pliocene
climatic optimum (about 3 to 4 million years B.P.)
corresponds to a warming of 3 to 4 DC." Considerable
effort has gone into studies of lake levels, types of
vegetation, and so on during those eras for indications of
what a warmer world might entail. However, Crowley
(1990) has concluded that geologic records yield no
completely satisfactory analogs of global warming because,
among other things, the earth's orbit around the sun is
different now than in the past, and nothing in the geologic
record matches the current rate of increase in
concentrations of greenhouse gases.

Although paleoclimatic data do not yield exact
analogs of future climates, they are valuable in
investigations of the physical mechanisms that control
climate. Larius et al. (1990) have calculated that between
40 and 65 percent of the variance in the Vostok
temperature record can be explained by variations in the
concentrations of CO2 and CH4 during the same period,
and about 80 percent by a combination of orbital forcing
(the Milankovitch effect) and greenhouse gas
concentrations. They also employed multivariate analysis
to deduce that the feedback factor that applied to
temperature variations driven by variations in the
concentrations of CO2 and CH4 over the 150,000 years
before the present was somewhere around 3. This result is
intriguing because it agrees fairly well with the theoretical
calculations of Ramanathan (1981). It is also in reasonable
agreement with recent output from general circulation
models, which will be discussed next.

5. PREDICTIONS FROM GENERAL
CIRCULATION MODELS

The results of Ramanathan (1981) were based on a
one-dimensional ocean-atmosphere model. The actual
ocean-atmosphere system is three-dimensional, which
allows the transport of heat from place to place by winds
and ocean currents. The only models capable of handling
the three-dimensional system are the general circulation
models (GCMs). There is an extensive literature
describing the various GCMs developed at different
institutions and providing sample results; space does not
permit any description of the models in this paper.

Most of the simulations of global warming using
GCMs have modeled a world in which the atmospheric
concentration of CO2 is steady near 600 ppm,
approximately double that of the year 1900. The results
indicate that doubling CO2 will raise the average surface
temperature of the earth by some amount in the range of
1.5 to SoC. Recent estimates are clustering around 3 to 4
dc. Most of the models indicate that warming in the polar
regions will be more pronounced than warming at lower
latitudes. One comforting fact is that the changes in
precipitation predicted by the models are within the range
of the rainfall fluctuations that have occurred from natural
causes over the past few hundred years. Unfortunately the
models do not simulate natural precipitation patterns very
well, and they do not agree concerning regional impacts on
either temperature or precipitation (Grotch and
MacCracken 1991).

L

Most doubled-C02 model runs simulate an earth
whose climate has stabilized following a sudden doubling
of the atmospheric CO2 concentration, with no further
increase thereafter. Such model runs do not correspond to
the actual earth, with its steadily increasing concentration
of greenhouse gases. Doubled-C~ scenarios can be
viewed as rough estimates of a condition through which the
earth may pass on its way to even higher concentrations of
greenhouse gases. As noted in section 2, the combined
effect of increases in C~, CH4, and CFCs is likely to
produce the equivalence of a doubled-C02 atmosphere
around 2050, but the onset of global warming will be
moderated by the tremendous thermal inertia of the ocean.
As some recent papers put the time constant for adjustment
of the temperature of the ocean's mixed layer to a thermal
forcing function at roughly SO years, it appears that the
theoretical predictions of climate for a doubled-C02 world
should be most applicable to some time around 2100.
However, upwelling of cold, deep-ocean water that has
never been exposed to an enhanced greenhouse effect will
continue for centuries and prevent the establishment of a
true equilibrium condition.

Critics of GCMs often refer to their coarse spatial
resolution, with grid points typically separated by a few
hundred kilometers in the horizontal, which precludes
representation of mesoscale and local phenomena except
through parameterization. However, the crude
representation of physical processes in the models is
probably a more serious deficiency. At this time, the key
physical uncertainties center around ocean-atmosphere
interactions, the radiative effects of clouds, and the albedos
of surfaces covered by ice or snow. As all Ihree factors
involve the possibility of significant feedback effects, their
correct representation is of critical importance.

The first ocean-atmosphere GCMs which modeled
the ocean at all treated it as a swamp, that is, a wet surface
with no capacity to store heat. More recent versions (e.g.,
Meehl and Washington 1990) treat the ocean as consisting
only of its mixed layer. That allows the ocean to store
heat. Models that treat exchanges of heat between the
mixed layer and the ocean depths are just beginning to
appear.

Clouds are sites where latent heat is transformed
into sensible heat (and sometimes vice vl~rsa), and they are
important to the earth's radiation budget (Fouquart et al.
1990; Arking 1991). Because the dimensions of individual
clouds are much smaller than a typical GCM grid spacing,
clouds must be parameterized in the models, and several
parameterization schemes are in use (Randall 1989). A
recent intercomparison of 19 different GeMs concluded
that most of the variation in results was attributable to
differences in the models' depiction of cloud feedback
(Cess et al. 1990).

Clouds simultaneously cool the earth by scattering
solar radiation and warm the earth by intercepting IR from
the ground and reradiating some of it downward as well as
to outer space. The relative importance of the two effects
varies with time of day, latitude, and cloud characteristics.
Important cloud characteristics in this regard include

112