Rewriting Arctic History
Posted by Jeff Condon on September 8, 2009
Lucy Skywalker, asked to put up a link to the graph below Fig 2. She sent me a short post in her own words which I think do a good job of catching one of the issues of the recent Arctic temperature reconstruction. Her work was done in response to the recent paper by Kaufman et al. which alleges that temperatures steadily declined for 2000 years before suddenly spiking. Lucy’s post compliled temperature records from John Daly’s work for the region all around the Arctic frim GISS and HadCRUT for visual comparison to the hockey stick curves presented by the Kaufman paper.
First and likely unsurprisingly for a proxy paper, I’m very skeptical of it. In fact it’s hard for the engineer in me to imagine the hubris of its conclusions considering it uses unverified proxies and thoroughly debunked mathematical methods but it will probably be difficult to disprove because it stands on the foundation of two dozen other papers filled with their own proxy mathemagic. Remember, I was far less strong worded about Steig09 on first blush. I’m waiting until I have time to get all the data from the original sources and have tried already to convert the jpg data in the article itself for verification. So far it has been made a particularly difficult road by the authors methods of publication. It would be better if people other than myself contacted the authors to request data and code because my Id has a reputation but I will be doing it in time. Anyway, I have a great deal of comfort with the fact that this particular temp reconstruction is equally as related to temperature as it is to magnetic flux trends on Neptune and the coarseness of cecropia moth antennae.
Figure 1 is a plot of the reconstruction having the same results as the Mann08 (bottom graph). Note the lack of MWP, no Little ice age and basically no variance other than the shock and recovery so easily created by bad data sorting mathematics.
Lucy did the right thing and went to find the actual measured temperature data creating a cool link to allow people to click on different temp curves and see the historic data of the actual stations. The actual data from uncorrected sources is critical for understanding whether the result is reasonable.
We know HadCRUT and GISS both provide substantial corrections to the ground temp datasets, in my opinion these are very likely overcorrected. I could do a whole post on how critical exaggerated surface station temp trends are to the correlation coefficients in the paleo papers. Advocates who hold themselves out to be scientists have reason to fear the corrections that surfacestations by Anthony Watts may reveal and we have seen some early mocking even from organizations such as the NOAA. I don’t have inside information on this, but it would make a huge difference in climatology to see even a small reduction of trend simply due to the difficulty of achieving unprecedented results with moderate correlation values.
Getting back to the story Lucy wrote this:
John Daly (1943-2004) was a superb climatologist and skeptic of AGW. His website “Waiting For Greenhouse” is still excellent and is occasionally updated by supporters. In particular, I never forgot his friendly collection of temperature records. He used the best of the oldest records, particularly Arctic records, that often go back well into the 1800′s; he also avoided those with UHI where possible – except to compare neighbouring records to demonstrate the distortion of UHI. He was before the time of Anthony Watts and the Surface Stations project, but he made available the evidence early on in a highly iconic way.
So when Kaufman et al published yet another unspeakable hockey stick, this time purporting to show Arctic temperatures rising dramatically in the twentieth century, I thought, I don’t remember a single one of Daly’s Arctic temperature records showing anything like that. But like a good scientist I checked – and still saw nothing. I then thought that we need to see ALL those worthy Arctic temperature records together on one page, “circling the Arctic” so that there is no loophole for accusations of cherrypicking.
Again she has the right idea, look at the data and what do you see. Unfortunately in this paper, other than choosing HadCRUT the cherry picking has to do with the leaving out of some proxies and use of other known hockey stick proxy series ‘Yamal’ which require very little effort or knowledge on behalf of the authors. They then simply use HadCRUT which has the highest known trend of any surface temp measurement for scaling through a CPS data masher. The whole process is very simple to manipulate but as it is built on a foundation of two dozen other ‘preferred’ results, it is more difficult to disprove.
Lucy then requested a further audit of these results. I would also like to see an audit. It is a soon to be run project which shouldn’t be time consuming to replicate if the data was easily available. My strong impression is that this paper will come down to the data chosen rather than the math or methods of the final result. How can we then make an argument against the paper, what do we say? – the scientists were biased in their source data choice?
Check out the link at Lucy’s site by clicking on the figure above. She’s set it up so any of the individual temperature graphs can be viewed full size by clicking on it.