The proof of climate change

Where is the proof that climate change is real and that humans are at fault? Climate sceptics love to pounce on the fact that scientists cannot prove that humans are the cause of climate change. Tony Abbott has said previously that climate science is “absolute crap”. More recently he has clarified that he meant to say the science of climate change is not yet “settled”. These comments presumably refer to climate scientists’ inability to prove that the climate is warming and that humans are contributing to it.

The concept of “proof” in science relates to the way in which scientists form and test hypotheses. But, as discussed by Massimo Pigliucci in his book Nonsense on Stilts, the way scientists go about testing these hypotheses depends on the complexity of the system they are trying to explain.

For example, a chemist in the laboratory may be trying to determine whether some compound “A” causes a particular reaction “B” to occur. To test this hypothesis, the chemist would probably set up an experiment to test whether reaction B occurred in the presence of A, and compare this with what happened in the absence of A. The chemist can follow this particular line of inquiry because they have the ability to manipulate and control most or all of the different factors that could affect the outcome of the lab experiment. An experiment is highly repeatable in controlled conditions, meaning that if the experiment is repeated over and over, the same results can be obtained consistently. Thus, predicting the behaviour of a system becomes very accurate when most of the variables are accounted for.

But imagine now that the chemist now has to perform their experiment out of the lab. The temperature and light intensity now fluctuates, and a little dust or dirt gets into the reaction. A little of compound A blows away in the wind before the chemist can add it to the reaction. Reaction B does not occur. The frustrated chemist repeats the process and this time reaction B does occur. Another repeat of the experiment produces a negative result. Once a system moves outside the lab and into the real world where some variables cannot be controlled, the system becomes less predictable.

Now consider an atmospheric scientist, who, like the chemist, wants to know whether compound A causes reaction B to occur in the atmosphere. In this situation, the atmospheric scientist cannot perform an experiment to change the concentration of compound A in the air - that would be impractical and irresponsible. Instead, the atmospheric scientist has observed that the concentration of compound A has increased in the atmosphere over the last few decades. Furthermore, the rate of reaction B has also increased during the same time period. Is the atmospheric scientist able to confirm or refute their hypothesis that compound A causes reaction B to occur in the atmosphere? Given that an experiment in the atmosphere is not feasible, is the current evidence sufficient to support the hypothesis? Or is the question just unanswerable? Is the science just “crap”?

Not all streams of science are able to test hypotheses by conducting a controlled experiment. While the outcome of a controlled experiment is highly repeatable, the results of the experiment have limited relevance in the real world, where conditions are impossible to standardise. On the other hand, it is difficult to conclusively demonstrate cause-and-effect relationships in natural systems because there are so often multiple factors that determine the outcome. Streams of science such as ecology and climate science often rely on making observations to identify trends and links between potential causes and effects. But scientists in these fields must report these results using sufficiently cautious language, using phrases such as “the evidence suggests…” and “our results may mean…”

Pigliucci argues that this does not make the science “crap” but instead reflects the limitations that scientists face in answering questions about complex systems. Unless we have a spare planet earth that we could observe, subject to exactly the same conditions as our own, except devoid of human life, proving in a scientific sense that climate change is human-induced is an impossible task and a foolish endeavour.

Scientists studying these complex systems are stuck between a rock and a hard place when it comes to communicating their work. The philosophy of science is such that definite conclusions are made only when hypotheses are tested in controlled experiments. The strict peer-review process, in which published science research is scrutinised by other experts in the field (I’ve discussed this in a previous post), effectively discourages scientists from making outlandish, unfounded claims. In contrast, those in business and politics want to see strong definitive results from science research, particularly when deciding whether to make a monetary investment in a discovery or when making government policies based on the results of a study.

But in some situations, as is the case with climate science and climate change, another aspect needs to be taken into account. With the climate data we have and the trends we have identified, we need to weigh up whether the price of inaction outweighs our need for a controlled experiment to prove our hypothesis. With this in mind, and considering that a controlled experiment to test the hypotheses is not feasible, I think we need accept that we are as close to proof as we are going to get, and we to act now.