First let us deal with the inevitable question that comes up when one addresses this topic - what makes you think you are qualified to criticise computer models?

Your bending author:
  • Started digital computer modelling in about 1960 on the first digital computer delivered in the UK for academic research, after wasting a year on an analogue computer. The machine was a Ferranti Pegasus, which had considerably less computing power than a modern hand-held device. Indeed, the computer you are using to read this almost certainly has more computing power than the whole world had then.
  • In the subsequent forty years, reviewed hundreds of computer models - in undergraduate reports, PhD theses and, as a consultant, in industrial applications.
  • Formulated the law of computer models long before global warming hit the headlines.
Most computer models are nonsense. This does not include those used by engineers in designing airplanes, bridges etc., which are based on detailed experiments on the systems involved and tested in a variety of real conditions before being used.

The reason they are nonsense is that they tend to be based on guesses of the value of coefficients assumed, particularly and disastrously feedback coefficients. There are few, however, that are quite as bad as climate models, where the physics of the interactions between variables and parameters is virtually unknown to mankind.

So, in inevitable synchrony with the issue of the Obama Government's excuses for draconian taxation, we have the UK media filled with dire predictions. The Times, The Telegraph and, of course, the BBC gave copious coverage, with many supplementary articles. The source is the Met Office, one of the running jokes in the few pubs remaining after the New Labour revolution. Not only have they got the annual forecast grotesquely wrong two years running (having forecast great heat, presumably on religious grounds) but they are not very good at telling us what the weather is going to do tomorrow. Their third attempt at predicting a "barbecue summer" is not going very well so far, but they are bound to get it right one day.

Not to worry, however, because they now have an outrageously expensive super-computer. Unfortunately, if you write a program to print "two plus two equals five" you get the same result however super the computer.

Imagine you settled down in your seat in a jumbo jet and noticed a plaque on the back of the seat in front which reads "This machine was designed with the aid of a super-computer. We did not know the values of all the parameters, so had to guess most of them." You would get off in a hurry. Yet the world's political and media establishment are asking you to gamble the economic future of yourself and your descendants on just such a proposition.

Computer modelling is one of the most powerful, yet dangerous, tools available to mankind. To be useful it has to be hedged around with checks, tests and precautions. Ruthless, politically-motivated, members of the new establishment are not concerned with such niceties. They want your money and your acquiescence and if it takes a (to say the least) dubious computer model to get them, so be it.