Geoscience Reference
In-Depth Information
twelve years to start reversing the rate of
greenhouse gas emissions before the global
changes become irreversible. Again the un-
certainties make such an assertion very
tenuous.
A different type of problem with math-
ematical modeling is illustrated by the
ipcc's aforementioned failure to include
melting ice sheets in its projection of sea
level change by the turn of this century.
This failure occurred because the panel said
there were too many uncertainties to make
a projection. It was a case of overdepen-
dence on models: since the panel couldn't
model the effect of melting ice sheets, it
would not make any projection at all. Oth-
ers, mostly on state and national sea level
rise panels, have had to roughly estimate
the contribution of the ice sheets based on
field and satellite observations.
Claims that a model is valid because it
has successfully reproduced a past event
(hindcasting) should also be viewed with
skepticism. As pointed out by Naomi
Oreskes of the University of California, San
Diego, hindcasting assumes that the com-
plex natural systems being modeled will be
subject to the same forces and will respond
the same way all the time—past, present,
and future.
For example, the model genesis is used
by coastal engineers to predict, among
other things, the amount of beach erosion
in the future. The model is “calibrated” by
hindcasting the erosion during some previ-
ous time span. The model is “adjusted” so
it comes up with the right answer and then
is applied to the future. But among other
problems, beach erosion is very sensitive
to storms; their intensity, duration, and
the direction from which they come are
variable factors that contribute to chang-
ing erosion patterns. To apply the model to
the future requires the unlikely assump-
tion that the schedule of storms in the fu-
ture will be quite similar to the schedule of
storms in the past.
With regard to predictive mathematical
modeling, the ever-colorful James Love-
lock argues, “we tend to be too hubristic to
notice the limitations. If you make a model,
after a while you get suckered into it. You
begin to forget that it's a model and think
of it as the real world. You really start to
believe it. We really don't know what the
clouds and aerosols are doing. They could
be absolutely running the show.” As a result
of our belief in models we are minimizing
direct observational data.
Jumping on bandwagons
Research scientists tend to be a skeptical
lot, but a couple of human-nature hazards
facing them can dull the sharp edge of sci-
entific cynicism that characterizes good
science. One of these hazards is the “band-
wagon” effect and the other is the “state of
siege” effect. The bandwagon effect occurs
when the vast majority of scientists sup-
port some idea such as some element of
global change. Their support for the idea
can cause the less venturesome and less
