Thomas Fuller on polar-bear-gate at Cliscep

This is an extended version of a comment made at Thomas Fuller’s cliscep article Okay, just one more post on polar-bear-gate… I promise…

There are three things highlighted in the post and the comments that illustrate the Polar Bear smear paper as being a rich resource towards understanding the worst of climate alarmism.

First is from Alan Kendall @ 28 Dec 17 at 9:35 am

But what Harvey et al. ignores is that Susan Crockford meticulously quotes from the “approved canon of polar bear research” and exhorts her readers to read it (making an offer to provide copies of papers difficult to obtain). She provides an entree into that canon- an entree obviously used by many and probably to the fury of polar bear “experts”.

This is spot on about Susan Crockford, and, in my opinion, what proper academics should be aiming at. To assess an area where widely different perspectives are possible, I was taught that it is necessary to read and evaluate the original documents. Climate alarmists in general, and this paper in particular, evaluate in relation collective opinion as opposed to more objective criteria. In the paper, “science” is about support for a partly fictional consensus, “denial” is seeking to undermine that fiction. On polar bears this is clearly stated in relation to the two groups of blogs.

We found a clear separation between the 45 science-based blogs and the 45 science-denier blogs. The two groups took diametrically opposite positions on the “scientific uncertainty” frame—specifically regarding the threats posed by AGW to polar bears and their Arctic-ice habitat. Scientific blogs provided convincing evidence that AGW poses a threat to both, whereas most denier blogs did not.

A key element is to frame statements in terms of polar extremes.

Second, is the extremely selective use of the data (or selective analysis methods) to enable the desired conclusion to be reached. Thomas Fuller has clearly pointed out in the article and restated in the comments with respect to WUWT, the following.

Harvey and his 13 co-authors state that WUWT overwhelmingly links to Crockford. I have shown that this is not the case.

Selective use of data (or selective analysis methods) is common on climate alarmism. For instance

  • The original MBH 98 Hockey-Stick graph used out-of-date temperature series, or tree-ring proxies such as at Gaspe in Canada, that were not replicated by later samples.
  • Other temperature reconstructions. Remember Keith Briffa’s Yamal reconstruction, which relied on one tree for the post-1990 reconstructions? (see here and here)
  • Lewandowsky et al “Moon Hoax” paper. Just 10 out of 1145 survey respondents supported the “NASA faked the Moon Landings” conspiracy theory. Of these just 2 dogmatically rejected “climate”. These two faked/scam/rogue respondents 860 & 889 supported every conspiracy theory, underpinning many of the correlations.
  • Smoothing out the pause in warming in Risbey, Lewandowsky et al 2014 “Well-estimated global surface warming in climate projections selected for ENSO phase”. In The Lewandowsky Smooth, I replicated the key features of the temperature graph in Excel, showing how no warming for a decade in Hadcrut4 was made to appear as if there was hardly a cessation of warming.

Third, is to frame the argument in terms of polar extremes. Richard S J Tol @ 28 Dec 17 at 7:13 am

And somehow the information in those 83 posts was turned into a short sequence of zeros and ones.

Not only one many issues is there a vast number of intermediate positions possible (the middle ground), there are other dimensions. One is the strength of evidential support for a particular perspective. There could be little or no persuasive evidence. Another is whether there is support for alternative perspectives. For instance, although sea ice data is lacking for the early twentieth-century warming, average temperature data is available for the Arctic. NASA Gistemp (despite its clear biases) has estimates for 64N-90N.

The temperature data seems to clearly indicate that all of the decline in Arctic sea ice from 1979 is unlikely to be attributed to AGW. From the 1880s to 1940 there was a similar magnitude of Arctic warming as from 1979 t0 2010 with cooling in between. Yet the rate of increase in GHG levels was greater from greater in 1975-2010 than 1945-1975, which was in turn greater than the period decades before.

Kevin Marshall