Modelling climate change


An investigation into the reported ‘fudging’ of figures at the Climate Research Unit (CRU) at the University of East Anglia has found no evidence of malpractice. So does this mean we should have complete faith in the research and models produced by this and similar bodies? I’m not so sure. The task of modelling and predicting climate change is a very difficult one. Earth’s climate, like many natural phenomena, is complex and requires some simplifications and assumptions in the modelling process. The language of the IPCC reports, however, doesn’t seem to take into account the likelihood of errors in climate change projections; this is misleading on their part.

It could be said that in this situation, given the apparent severity of the possible outcomes, it is best to ‘err on the side of disaster’, as it seems the IPCC does. This is an argument for a type policy response though; scientific research should never be skewed in such a way.

I would say that my biggest worry about the climate change models is related to measurement errors. Owing to the nature of the subject under observation, measurement errors are likely. It is not often pointed out that projecting such observations into the future then compounds this measurement error even further; in the jargon, this is called ‘sensitive dependence on initial conditions’, the initial conditions being measurements such as temperature. Given the sensitivity of the models to even tiny differences in the initial measurements, the projections can be vastly different because of a tiny measurement error to start with. Just like the proverbial butterfly in India that can help cause a hurricane in Oklahoma, measurement errors can impart massive distortions in the projections of variables such as future temperature and sea level.

I’m not a climate change sceptic, but a climate model sceptic. We just don’t understand the complex systems at work as well as the IPCC seem to think we do.