Environmental Engineering Reference
In-Depth Information
we really want to see is the best prediction of what will
happen in the future, but we cannot know which of the
many models agrees best with future facts until enough
time has passed to collect the necessary data to show the
differences between reality and model prediction.
Remember that all of the models say the temperature
is going up; they differ by how much it will go up (and
how fast). As we shall see, it will take about
years for
the temperature to change enough to show which of the
models does the best job. If it turns out that the predic-
tions giving the larger temperature rise are the correct
ones, it will be much more dif
cult to limit the ultimate
temperature increase if we do nothing until we know
the answer.
I will use the A
scenario to make an estimate of how
long it will take to learn which of the models are most
nearly correct. When enough time has passed the models
will not need scenarios to compare with the data. They
will have the actual change of greenhouse gas concen-
tration over time as their input. Scenarios will still be
needed to predict the future under different assumptions
on energy use, technology development, population
growth, etc., but the spread in the future temperature rise
will be reduced because those models that disagree with
the observations will be discarded (or more likely, their
knobs readjusted).
The complicating factor is noise in the system. The
average temperature of the Earth varies randomly from
year to year by an average of
F) above
and below the trend line. Sometimes the random jump is
less, sometimes more, but the long-term average noise is
quite consistent.
Figure
C (about
.
.
.
for scenario A
shows bands
