'Fatigue' in the climate system
In 1954, two de Havilland Comets broke apart in flight, leading to lives lost and the suspension of service for what was then a state-of-the-art aeroplane.
The cause was traced back to metal fatigue, a sudden failure whose underlying cause was the slow weakening of the material caused by repeated cycles of loading.
Before and since, scientists have studied metal fatigue, to understand its basic causes and to discover if there are tell-tale signs that signal impending collapse.
These studies have intimately involved the use of computer models.
Although the causes are far better understood than half a century ago, precise prediction of catastrophic failure remains impossible - as it does with earthquakes.
Could the same be true of the Earth's climate system?
The best computer models of climate, he says, are not able to "predict" at least four major transformations in the past:
- the Paleocene-Eocene Thermal Maximum, when temperatures 56 million years ago rose rapidly
- the sharp drying of North Africa 9-5,000 years ago
- serial collapses of the overturning circulation in the North Atlantic (in common parlance, "the Gulf Stream shutting down")
- sharp warmings - Dansgaard-Oeschger events - recorded in Greenland ice cores.
If computer models of the hugely intricate and interconnected climate system cannot predict those, how confident can we be that they will be able to predict something similar, if it is coming, in the near future?
And if we cannot be confident, what does that mean in terms of the choices that societies make - over "insurance" measures such as flood defences, and over how stringent to be on curbing carbon emissions?
In the acrid climate blogosphere there are many commentators who would agree with Professor Valdes' contention about lack of confidence in computer models.
Their conclusion, typically, is that society should not take any steps to mitigate emissions until the projections are surer.
Going back to the analogy of aeroplanes, this is tantamount to arguing that it's fine to get on board any craft unless it's been shown to be unsafe.
And this is where the sceptical commentators would part company from Paul Valdes; because he draws the opposite conclusion.
"We need to be cautious," he writes. "If anything, the models are underestimating change, compared with the geological record.
"According to the evidence from the past, the Earth's climate is sensitive to small changes... simulations of the coming century with the current generation of complex models may be giving us a false sense of security."
How often the professor travels by plane I'm not sure, but I sense he wouldn't be prepared to get on board unless he was confident it had been shown to be safe - a radically different approach from the sceptics.
The aeroplane analogy isn't exact, for a number of reasons - the simplest being that if something like the Comet crashes happens, engineers can look back, understand it, and put it right - which may lead to a better 'plane as well as better models.
This is a luxury we don't have with the climate system; as has often been observed, there is no Planet 2.0.
On the other hand, the human spirit thrives in an atmosphere free from unmerited fears.
So whose conclusion are you on board with?