This post by David Archibald at WUWT extolls the virtues of “wiggle matching,” then applies it — or so it claims — to match the CET (Central England Temperature) from 1703 through 1737, to that from 1978 through 2012. Archibald then notes that the 1703-1737 period was soon followed by the unusually cold year 1740. His conclusion? That we might be in for a similar year of extreme cold in 2015.
There are lots of problems with his efforts, but the main one is that the wiggles don’t match.
So what does that period up to 1740 wiggle-match with? It matches with the warmth of the last 30 years:
The graph above shows the Central England Temperature (CET) record from 1703 to 1745 as the blue line. Plotted on it is the CET record from 1978 to 2012. Normally when you align 34 year lengths of temperature records you don’t get any correlation. The correlation on this particular matchup is 0.112. The statisticians amongst us can argue over whether or not anything can be read into that. If something can be read into it, we only have to wait two years to experience the consequences.
First of all, when you align 34 year (or any) lengths of temperature records you will get some correlation. If the records are unrelated then the correlation is likely to be small (how small depends on the noise level and structure and the number of data), but it is unlikely (damn near impossible) to be zero. The question is, is the correlation large enough to be meaningful? If so, what does it mean?
Actually the 0.112 isn’t the correlation, it’s the squared correlation — and I get 0.111 rather than 0.112. The correlation itself is 0.333.
Here’s the data without the artificially curved lines:
Does such a correlation mean anything? Testing it, the p-value is 0.0508, which — just barely — fails statistical significance at the 95% confidence level. Ordinarily that might be interpreted as a sign that Archibald’s wiggle-matching might have some meaning or might not. But in this case, the answer is a definite no.
Why? Because Archibald went looking for correlation — even did “wiggle matching” to impose it — and the best he could come up with failed to achieve 95% confidence. If you’re trying to match up two data sets, and you still can’t get to 95% confidence, then you’ve got nothing.
But wait — there’s more! The correlation which is present (whether meaningful or not) isn’t because the wiggles match at all. It’s because the trends match:
When the trends are similar, as they are here, there’s going to be correlation. If we want to know whether or not the wiggles match, then we need to remove the trends first. That gives us this:
Now the correlation isn’t 0.333 (sqaured correlation 0.111), it’s a pathetic 0.0806 (squared correlation 0.0065). For a data set this size, that’s about as meaningless as it gets; testing for significance, the p-value is 0.645, which is a resounding No Way!
Bottom line: David Archibald tried to predict an upcoming cold year based on “wiggle matching” with past data, but even though he was trying, he couldn’t get the wiggles to match. That didn’t stop him from trying to forecast the consequences of his purported cold year in another post at WUWT.
Of course, this is just another amusing example of silliness masquerading as science on the WUWT blog. David Archibald tried to “wiggle-match” old CET data to modern CET data and utterly failed. But because he didn’t do any analysis of his result, he doesn’t even know that! Because he lacks the basic knowledge of how correlations can arise, and the difference between wiggles and trends, he doesn’t even know that the “wiggle” part of the correlation is vastly smaller than he estimates (and again, meaningless, which again, he remains ignorant of).
But, in my opinion, David Archibald isn’t really interested in understanding science, or what we can expect from future climate. He is, however, very interested in justifying a scenario in which we should be afraid of cold rather than heat, because his bottom-line goal is to sabotage any effort to take global warming seriously. Just my opinion.
Also my opinion: the most interesting thing about this sordid little comic episode is how it shows that the WUWT blog, always eager to declare how incompetent actual working climate scientists are, has such low scientific standards that they will actually publish this drivel. Just my opinion.