"The IPPC is chartered only to document human effects on climate and thus does not countenance any information or hypothesis that does not support their human causation hypothesis.
Since both the human‐impact and natural hypotheses use correlations to interpret causation, how can causation be resolved? The obvious answer is predictability. The validity of either hypothesis can be tested by natural experiment.
The AGW hypothesis predicts rising carbon dioxide concentrations causing increased global temperature, rapidly rising sea levels at rates greater than pre‐1900 rates, and massive loss of glacial and oceanic ice. There is no question that the concentration of CO2 is increasing.
Conversely, the solar or natural hypothesis predicts that temperatures should gradually decrease over the next thirty years before rising once more, a Gleissberg Cycle with typical slow cooling followed by rapid warming. Until the next glacial advance, perhaps in a couple of thousand years, sea levels are predicted to slowly rise until a full connection between the Arctic Ocean and the North Atlantic is established, opening the Arctic Ocean and giving rise to the moisture from which continental glaciation can re‐establish.
Nature’s grand experiment is underway. Earliest returns are that temperatures are not going up despite the rise in carbon dioxide. More years will be required to resolve the hypotheses. The predictions are the test.
Correlation does not imply causation."
[Illustrations, footnotes and references available in PDF version]
Correlation is the easiest method to assign causation for an event, even though it is the least valid method. Take an automobile race, for instance. The last four races were won by blue cars. Blue cars win races. That is an example of correlation driving causation. In reality, Vettel was driving the blue car. He just won the Grand Prix championship for 2013. The color of the car had nothing to do with winning the races. There are frequent correlations that do not identify the causes of events, but politics and media jump on the simplest correlations because they do not require extensive research nor complex analysis.