Additional info from Skeptical Science
This argument betrays a misunderstanding of the difference between weather, which is chaotic and unpredictable, and climate, which is weather averaged out over time. While you can't predict with certainty whether a coin will land heads or tails, you can predict the statistical results of a large number of coin tosses. Or expressing that in weather and climate terms, you can't predict the exact route a storm will take, but you can predict the average amount of rainfall a given region will get in a year.
Climate prediction is a difficult art that's constantly being refined. There's the problem that future behavior of the sun is very difficult to predict. Similarly, short- term perturbations like El Niño or volcanic eruptions are difficult to model. Nevertheless, climate scientists have a handle on the major drivers of climate.
James Hansen's 1988 climate predictionsWay back in 1988, James Hansen projected future temperature trends. Those initial projections show remarkable agreement with observations right to the present day (Figure 1).
Hansen's Scenario B (described as the most likely option and in hindsight, the one that most closely matched the level of CO2 emissions) shows close correlation with observed temperatures. There are deviations from year to year, but this is to be expected. The chaotic nature of weather will add noise to the signal but the overall trend is predictable.
Modeling the aftermath of the Mount Pinatubo volcanic eruptionWhen Mount Pinatubo erupted in 1991, it provided the opportunity to test how successfully models could predict the climate response to the sulfate aerosol injected into the atmosphere. The models accurately forecast the subsequent global cooling of about 0.5 degrees Celsius soon after the eruption. Furthermore, the radiative, water vapor, and dynamical feedbacks included in the models were also quantitatively verified (Figure 2).
Comparing IPCC projections to observationsThe study “Recent Climate Observations Compared to Projections” compared projections made by the IPCC in 2001 of global temperature change (colored dotted lines) with observations from HadCRUT (blue) and NASA GISS data (red) (Figure 3). The thin lines are the observed yearly average. The solid lines are the long-term trends, which filter out short-term weather fluctuations.
It's immediately apparent the IPCC underestimated temperature rise with observations warmer than all projections (but within the grey uncertainty area). The paper proposes several possible reasons for the difference. One is intrinsic internal variability which is possible over such a short period. Another candidate is climate forcings other than CO2, such as aerosol cooling, being smaller than expected.
A third candidate is an underestimation of the effects of amplifying feedback loops in the climate system, like the loss of Arctic sea ice or an increase in water vapor. You can learn more about the IPCC’s 2001 projections here.