As we were saying only last month, the motto du jour is get your rationalisation in first. The latest wheeze among the doomsayers is that hell fire is being postponed. Of course, it would have been more impressive if it had been published before the recent decade of measurements showing no warming at all. As it stands, it is nothing more than a testament to the infinite tunability of computer models. The warmers are getting more and more like those traditional predictors of the end of the world who, when the event fails to happen on the due date, announce an error in their calculations and a new date.

There is a new tendency that seems to afflict both believers and deniers alike, waiting for the next monthly data point and trying to make deductions and forecasts from it. A single new data point on a noisy graph tells you nothing about the development of trends (See the example of What happens next? in this presentation, slides  28-30). Fundamentals, such as the Uncertainty Principle and the estimation of trends ensure that a single point in time tells you nothing about the evolution of processes. Like the official definition of recession, you don’t know it has happened until after the event.

PhD training used to be about eliminating tendencies such as getting excited about a new data point and loading upon it all sorts of fantastical premonitions that it is unable to bear, but that was another age and besides the culture is dead. Furthermore, there is nothing more depressing for a referee than to find the introduction of a filter of mine own invention, as happened with the revival of the hockey stick by a pseudonymous author analysing data from a prominent Urban Heat Island. It is in the nature of the beast that at a certain stage in our development, scientists fall in love with the sheer romance of signal processing and rush to adopt the latest technique of linear algebra, including filtering, but it is a sign of maturity to stick to well known and fully understood procedures. In particular, a new procedure requires an enormous amount of analysis just to see what it actually does. More often that not (as for example in the introduction of the coppock) the only effect is to enhance the outcome desired by the author at the expense of perspicuity.

One of the most absurd examples of chartmanship you are ever likely to see is the one supposed to indicate an uprising of the dreaded methane. Not only are the units of the ordinates parts per billion, but the suppression of the zero is as dramatic as you can find. What is to all intents and purposes a horizontal line is transformed into a portentous curve, but even then it is indicative of a saturating phenomenon rather than an evolutionary one. It transpires, however, that the whole “threat” depends on a wriggle of noise at the very end of the plot. It might well transpire eventually that the curve is rising, but this nonsense does not establish it  by a long chalk.

Alas poor science!  Source

Average user rating