I had the profound joy of bumping into Mike Haseler on Twitter earlier today. After much bandying of words about global temperature the “debate” headed north to the Arctic. Here are the edited highlights:
I feel sure this story will run and run, but I’ve got to rush off up the M5 soon. Consider this as an appetiser, with the main course to follow in due course.
Watch this space!
9 thoughts on “Alternative Facts in the Arctic – Case Study 2”
If those are the highlights I would hate to see the lowlights. All he proved is that a howling jackal can call himself a physicist.
The “Scottish Sceptic” keeps on doubling down! I’ll go and update the edited highlights with this morning’s “debate” shortly. I’m sure this story will indeed run and run. The loquacious Mike reliably informs me via Twitter that:
Jim, to which temperature dataset was Mr Haseler referring in order to arrive at his “19 years without warming”.
I’ll do a couple of charts for you with the relevant trend(s). If it’s RSS, then it gives me a reason to bring the Moncktonite charts up to date.
I’ve been out and about meeting and greeting my new MP:
A surprisingly positive experience. Scott talks far more sense than Mel ever did! Getting back to the Scottish Sceptic, this is a close as I ever got to persuading him to part with the source of his data:
Jim, what is the story behind this graf Tony Heller is showing all the time. I understand the right side of it, it is from AR1. But where and when did Vinnikov state that Artic annual mean was 6” km2 i 1960?
Mr. Heller links to this 400+ page 1985 report from Lawrence Livermore National Laboratory:
“Projecting the Climatic Effects of Increasing Carbon Dioxide”
From page 152:
Other graphs bandied about during the “debate” were from the NSIDC:
and IPCC AR5, Chapter 4, p.326:
Your Scottish friend, Mr Haseler, clearly does not have the faintest idea how to interpret an R-squared value.
Used a similar technique to that employed when dealing with Monckton’s RSS claims last year, I could quite easily work out exactly what dataset and time window that Mr Haseler had used. His laughable claim is predicated on the period Jan 1998 – Jan 2017 of the UAH Version 6 Beta 5 dataset – so full-marks on the egregious cherry-picking front. (You might want to have a quick look at the analysis I did on the effect of the “upgrade” from Version 5.6 This appeared as Comment #17 on the ClimateGate 2 thread on the ASIF)
Using a linear trend line on this data does indeed yield an R-squared (or Coefficient of Determination) value of 0.0196
However, that number merely tells people (but not Mr Haseler, apparently) that a linear trend line is a bad fit. It most certainly does NOT tell the informed observer that there is no trend. It is possible to have an R-squared value of ZERO, even if there is a perfectly known non-linear relationship between the dependent and independent variables.
A quick example I knocked up was based on a pure sine-wave and a linear ramp function. The formula was simply…
Y = sinX + 0.00005X (where X is in degrees, 0 <= X <= 7575)
(NB No real reason for stopping at 7575. I was using Autofill, and happened to lift the mouse at that point, as it had by then gone through in excess of 20 cycles.)
The above formula yields a chart quite similar to the Keeling Curve, and could be projected beyond the end point with 100% precision.
If one is daft enough to apply a linear trend, the R squared value comes out at 0.0142 – somewhat below the value of 0.0196 which so raised Mr Haseler's ardour.
I'll write up some words and send you some diagrams tomorrow – I've been down the pub, and so am no longer capable of writing in the Queen's English.
However, even when somewhat sozzled, I can still take that sort of denier drivel to the cleaners.
I voted 19.75- 20.25, just so we all know where we stand ~;^o’///,<
Thanks for keeping the interwebs real Jim!
May I also say that I’m quite disturbed that a fellow scot also came up with the idea of Jevons Paradox…. hmmmmmmmmmmn- I think there’s something in that for all of us, ..don’t you!??!