Short-term Predictions - Climate Change: Identification and Projections - page 116

the eruption of Pinatubo in 1991 can be clearly distinguished. This effect
aside, the simulated average global temperature is essentially due to
anthropogenic forcing. The contribution of solar irradiance is imperceptible.
Figure 11.1. Short-term projections from the IPCC
The predictive power of these models is disappointing, if not non-
existent: in 2012, the real temperature was below the overall envelope of all
the simulations. It would clearly be wrong to criticize simulations started in
1986 for not reproducing the random fluctuations in later years, or the peak
in 1998, due to an El Nino episode which could not have been predicted in
1986, or the climate “pause” which followed. However, although the latter is
due to climatic fluctuations, the simulations should reproduce
similar fluctuations here and there, because of the oceanic chaos generated
by the simulations. This should have resulted in a bundle of projections with
a spread such that the actual temperature does not leave this range, as it did
in 2012.