Monday, November 30, 2009

Climategate gets worse

You should all know about the scandal that has been dubbed 'climategate' by now. Basically, a large amount of email and other data relating to the work of the Climatic Research Unit at the University of East Anglia has been leaked, and it does not present a pretty picture. We have conspiracy to evade Freedom of Information Requests by deleting data, schemes to freeze out academics who disagree with the CRU's position on global warming, and emails that suggest holes in the base data that supports Climate Change. However, the worst of it is that the computer models on which so much current political policy worldwide is based appears to be seriously flawed. Programmer notes suggest poor error handling, poor input data, and a number of hardcoded values in an effort to get the programme to match observed climate behaviour in the past. This raises doubts about the model's actual ability to predict the future. Complex computer models can quickly descend into complex garbage without good inputs and well-formed processing. Neither appears to be present here.

It appears that there has been very little effort on the part of the CRU to build decent software using proven programming and software engineering techniques. Instead we seem to have a very bright, but unqualified and inexperienced programmer churning his way through masses of computer code at a level of professionalism that can only be described as hobbyist. Knocked together programs are all right as one-off support for a Phd thesis or as part of an experimental cycle, but definitive climate modelling demands something better.

All of this leaves us with plausible hypotheses on global warming and climate change based on evidence that appears increasingly flimsy. Most seriously the CRU says that it has thrown away much of the base data on which their models are based, making some of their work unrepeatable and therefore unverifiable by anyone else.

Right now there is a broad consensus on climate change among the three main parties, based on what was supposedly solid scientific theories. If these are no longer solid, and if the public belief in global warming continues to slide, then this consensus will fracture.

Maybe it should.

No comments: