Short-term Predictions - Climate Change: Identification and Projections - page 124

which we will now carry out from the year 2000 should be just as inaccurate
as those of the IPCC.
11.5. A posteriori predictions
Firstly, we will make a prediction for solar activity from the year 2000
(see section 10.2). Figure 11.5 compares the prediction with the irradiance
actually observed until 2013.
Figure 11.5. Predicted and observed irradiance
Of course, our prediction method (see section 10.2) does not pretend to
compete with that of the solar physicists. It simply satisfies the requirement
of making a plausible short-term prediction, which owes nothing to
knowledge of a future which is as yet unknown.
For CO 2 , we adopted the RCP 4.5 and, for volcanism, a prediction using
the average value over the last millennium.
Figure 11.6 shows retroactive (or anticipated) predictions carried out
from the year 2000, and these projections bear no resemblance to any kind of
linear extrapolation. The observations do not leave the bottom of the bundle,
like in Figure 11.1, where the anticipation does indeed start later (2006).
In 1999, we were not yet looking into the climate issue, but had that been
the case, we would have been able to develop the identical identification
algorithms and predictions shown here. We would therefore have obtained