Thursday, July 12, 2007

IPCC: not even pretending to be scientific

Timmy has an interesting piece up at the Adam Smith Institute blog, based on a paper about scientific forecasting [PDF], written by Scott Armstrong and Keston Green.
Now, no, the authors do not claim expertise in climate science. What they do claim is expertise in how to make scientific forecasts. Indeed, without exaggeration, they quite literally wrote the book on the subject.

The essential point is that to be valid, forecasts must be more than just the expression of the scientists' thoughts or hunches, however dressed up they are in mathematics. There are so many points in climate models where judgement must be used (for the details of many processes are as yet unknown) that it is arguable whether these models are in fact "scientific forecasts" rather than simply exceedingly complex expressions of the thoughts and hunches. Which, as James Surowiecki has repeatedly pointed out, are subject to groupthink reinforcement, most especially if each supposedly independent voice is in fact communicating with all of the others.

Running part of the IPCC report through their long developed 140 point checklist the authors find that:
Of the 89 forecasting principles that we were able to rate, the Chapter violated 72.

Their final conclusions are summed up in these two lines:
We asked, are these forecasts a good basis for developing public policy? Our answer is “no”.

And...
We have been unable to identify any scientific forecasts to support global warming. Claims that the Earth will get warmer have no more credence than saying that it will get colder.

No, I have absolutely no idea whether the authors are right or not but it might be worth noting that Armstrong is backing his findings with a $20,000 bet on whether climate change will in fact happen or not. My real hope though is that someone actually does check and see whether they're right or not: after all, we're about to commit ourselves to spending trillions of dollars on the basis of the IPCC Report: shouldn't we actually find out if it is correct first?

Well, it would be a fucking start, eh?

4 comments:

Vindico said...

"of the 89 forecasting principles..the chapter violated 72" lol!!!!!!!!!!!
When will these little nerdy shits who call themselves politicians learn? Or more to the point when will the sheeple finally realise they have been brainwashed by the enviro-fascists and get a grip on reality?

bishophill said...

Kevin Trenberth has a piece up at Nature Climate reports in which he basically says "They're not forecasts, they're scenarios" and claims that this means that Armstrong's principles don't apply.

To use a suitably Devilish turn of phrase, this is utter bollocks.

John Nicklin said...

High time that someone provided proof for what we have known for a long time. The IPCC is a badly flawed organization with few, if any, redeeming values.

Predictions, forecasts, scenarios, who cares, they purport to tell us what the climate will be like in 100 years when they can't get next week right with 95% accuracy 19 times out of 20. In Canada our government weather people are no more accurate than a table of random numbers for 75% of the country. Why we pay them I don't know. But they bleat on endlessly about how were are all going to fry or drown.

Milo said...

It is apparently far easier to produce general predictions for large areas over long periods of time than it is to produce more specific predictions for small areas over short periods.
Well, so the chaps from the Met. Office tell me.