This paper is a must read for anyone who wants the truth about the dishonesty surrounding some of the original claims of climate change and the supposed "proofs" that it is an "existential threat".
Suffice it to say that there are a lot of Climate Alarmist Scientists who are not credible and their message of alarm is unjustified.
Here is an excerpt, which is only the tip of the iceberg.
----------------------------------------------
INTRODUCTION: MAKING THE MYTHS
It is now 10 years since the Climategate emails were released. The issues they raised continue to reverberate; even figuring in a decision last week of the United States Supreme Court to allow Michael Mann’s (US) defamation suits to proceed (see the dissent by Justice Alito), and in an August 2019 decision of the BC Superior court dismissing a similar suit (on which see more below). The immediate reaction at the time to the emails was visceral, even among “green” reporters, including George Monbiot as follows:
Pretending that this isn't a real crisis isn't going to make it go away. Nor is an attempt to justify the emails with technicalities. We'll be able to get past this only by grasping reality, apologising where appropriate and demonstrating that it cannot happen again
UK reporter Fred Pearce, who covered the story for the Guardian and who, unlike Oxburgh or Muir Russell, had actually read the emails, wrote in The Climate Files:
The evidence of scientists cutting corners, playing down uncertainties in their calculations and then covering their tracks by being secretive with data and suppressing dissent suggests a systemic problem of scientific sloppiness, collusion and endemic conflicts of interest, but not of outright fraud. (p. 241)
Given the importance of climate science in today’s society, all of us expect more of climate scientists than merely that they not commit “outright fraud.” Exoneration at such a low threshold would be small exoneration indeed.
However, rather than confronting the corruption and misconduct apparent throughout the Climategate emails, the climate academic community shut their eyes to the affair, eventually even persuading itself that the offending scientists were victims, rather than offenders.
This re-framing was made possible by numerous myths propagated about the affair, of which the following were especially important:
Myth #1: The Climategate scandal arose because “cherrypicked” emails were taken “out of context”.
Myth #2: The Climategate correspondents were “exonerated” following “thorough” and impartial investigations.
Myth #3: Scientific studies subsequent to Climategate have “confirmed” and “verified” the original Mann hockey stick.
These are only the major myths from a veritable tsunami of disinformation from the academic community. The myths are untrue and, in this article, we will explain why.
Yamal: Climategate in a Nutshell
Pretending that this isn't a real crisis isn't going to make it go away. Nor is an attempt to justify the emails with technicalities. We'll be able to get past this only by grasping reality, apologising where appropriate and demonstrating that it cannot happen again
UK reporter Fred Pearce, who covered the story for the Guardian and who, unlike Oxburgh or Muir Russell, had actually read the emails, wrote in The Climate Files:
The evidence of scientists cutting corners, playing down uncertainties in their calculations and then covering their tracks by being secretive with data and suppressing dissent suggests a systemic problem of scientific sloppiness, collusion and endemic conflicts of interest, but not of outright fraud. (p. 241)
Given the importance of climate science in today’s society, all of us expect more of climate scientists than merely that they not commit “outright fraud.” Exoneration at such a low threshold would be small exoneration indeed.
However, rather than confronting the corruption and misconduct apparent throughout the Climategate emails, the climate academic community shut their eyes to the affair, eventually even persuading itself that the offending scientists were victims, rather than offenders.
This re-framing was made possible by numerous myths propagated about the affair, of which the following were especially important:
Myth #1: The Climategate scandal arose because “cherrypicked” emails were taken “out of context”.
Myth #2: The Climategate correspondents were “exonerated” following “thorough” and impartial investigations.
Myth #3: Scientific studies subsequent to Climategate have “confirmed” and “verified” the original Mann hockey stick.
These are only the major myths from a veritable tsunami of disinformation from the academic community. The myths are untrue and, in this article, we will explain why.
Yamal: Climategate in a Nutshell
A good illustration of the three myths is provided by the Yamal story. In many ways it was at the heart of Climategate, yet very few commentators picked up on its centrality. Most tree ring temperature proxies do not have a hockey stick shape. A few do: some bristlecone pine records from the US (on which Mann’s hockey stick depended) and a few larch records from the Yamal peninsula of northwest Russia. The Yamal record was the key ingredient for virtually all the supposedly independent confirmations of the Mann hockey stick.
The first email in the Climategate archive, dated March 6 1996, was from a Russian scientist (Shiyatov) to Briffa requesting money to support their efforts to collect more tree ring data from the Yamal peninsula. One of the last Climategate emails (dated October 5, 2009, just a month before the Climategate release) was from scientist Rashit Hantemirov to a UEA colleague asking for advice on how to respond to a Finnish journalist who was investigating Briffa’s use of the Yamal data, based on findings Steve had published at Climate Audit. Yamal was, in fact, the mostrepeated theme in the emails, even though it never captured public attention. The emails provided extensive context for a controversy that had long been raging.
Since 2005, Steve had regularly criticized Briffa and the CRU for concealing an updated version of a proxy record from the Polar Urals which, unlike the original published in 1995, showed a strong Medieval Warm Period. The cold medieval segment of the Polar Urals series was critical to a few early hockey stick-like reconstructions, so much so that using the revised series would have overturned the original conclusions.
Briffa resisted disclosing the updated Polar Urals data and Steve only obtained it after a lengthy dispute with Science magazine. When it finally became available, the CRU scientists promptly dropped it from their studies and substituted one from the nearby Yamal peninsula instead. Whatever the stated reasons for doing so (and at the time none were given), the effect was to remove a proxy that now had a medieval warm period and replace it with one with a very strong hockey stick shape, especially due to a big jump after 1990.
The two series (Yamal and updated Polar Urals) gave contradictory information about the climate of the region in the medieval era, something not disclosed to readers of the very studies and reports that placed great emphasis on the importance of being able to make precise claims about the relative warmth of the medieval era. Over an 8-year period Briffa used the Yamal series repeatedly in his papers, but never published the data. Steve’s various requests for the data were ignored, but in 2008 Briffa published a study based on Yamal in a journal (Philosophical Transactions of the Royal Society) that had adopted strict data disclosure rules. After Steve asked the journal for the data the editor demanded Briffa provide it to Steve.
What immediately became apparent was that the post-1990 jump was based on a sampling flaw: the sample size collapsed at that point to fewer than the minimum number of trees required and the series should have been terminated prior to 1990. Also, Briffa had not used a large set of nearby tree cores that would have allowed the full interval to be covered, but doing so would have yielded a different overall profile, one with no hockey stick shape. Although numerous postMann hockey stick studies relied on the Briffa Yamal series to provide a supposedly independent confirmation of the Mann result, in 2013 the CRU quietly abandoned the Briffa Yamal series and substituted one based on a larger sample that looked a lot like the one Steve had computed back in 2009 by combining Briffa’s data with the larger nearby data set.
None of this misconduct was dealt with by the various inquiries. The Oxburgh panel ignored it entirely (see below). In Steve’s evidence to the Muir Russell panel (on which also see below) he showed the published version of Briffa’s Polar Urals series, with its cold Medieval series, and the updated resampling in which the medieval era was now very warm by comparison to the present. He also discussed Briffa’s secrecy and refusal to publish his data, which thwarted discovery of the weaknesses of his temperature reconstructions. The Muir Russell panel dismissed all these concerns on the basis that they were not published in academic journals. This was ridiculous reasoning since, first, much of the battle involved getting the journals to enforce their own data disclosure policies but this typically does not lead to an article in the journal, and second, by refusing to disclose the data Briffa was making it impossible for papers critical of his analysis from being published.
MYTH #1: THE EMAILS WERE TAKEN OUT OF CONTEXT
Climate academics repeatedly assert that the emails were taken “out of context” to create controversy, yet the reality is the exact opposite: the controversies already existed; the emails provided the disquieting context that exposed the depth of malfeasance. The most notorious emails (e.g. “hide the decline”, “dirty laundry”) concerned issues and controversies which had already been raised at “skeptic” blogs (especially Climate Audit). The emails provided background detail which was then analysed extensively in contemporary blog posts at Climate Audit. Rather than coming to terms with the revelations, the climate community has simply chanted “out of context!”, but never demonstrated that there exists an alternative context in which the emails were less damning.
We will show this for several of the most prominent and controversial emails, but the potential list is far more extensive.
Example 1: “Dirty Laundry”
In a remarkable 2003 email, which was discussed in detail at Climate Audit within the first two weeks of Climategate, Mann sent some undisclosed calculations from the Mann et al 1998 temperature reconstruction to Tim Osborn, a “trusted colleague”, telling Osborn that the series were his “dirty laundry” and needed to be kept strictly confidential so that they didn’t fall into the wrong “hands”.
The email sounds bad enough on its face, but, in context, it is even worse. Nor was Mann’s withholding of data an issue that originated with the Climategate emails: it was a longstanding controversy to which the Climategate email added additional and disquieting context. It also touches issues which, due to Mann’s libel lawsuits, linger on to this day.
Reconstruction Residuals
The email sounds bad enough on its face, but, in context, it is even worse. Nor was Mann’s withholding of data an issue that originated with the Climategate emails: it was a longstanding controversy to which the Climategate email added additional and disquieting context. It also touches issues which, due to Mann’s libel lawsuits, linger on to this day.
Reconstruction Residuals
The “dirty laundry” data series are called residuals. They are the differences between the proxy reconstruction estimates of past temperature and observed temperature records during the model estimation (“calibration”) and testing (“verification”) periods. Since the residuals measure the goodness-of-fit of the model, they are essential for computing verification test scores. In this email Mann was supplying residuals for reconstructions (which he grandiosely calls “experiments”) based on the post-1000, post-1400 and post-1600 intervals. The first two were critical since they determine whether it is legitimate to do the reconstruction back that far.
Numerous statistical authorities, including those1 cited in Mann et al 1998, recommend testing reconstruction validity using several different scores based on the residuals. Mann stated in his 1998 paper that he had computed two such scores, the Reduction of Error (RE) statistic and the r2 score. But in his paper and in the accompanying archive he only listed the RE values. He had not (and has never) released the r2 scores. Nor could they readily be computed from information disclosed with the original publication because, contrary to widespread belief among climate scientists, Mann’s archive omitted the complete reconstructions for each time step. For the signature Northern Hemisphere (NH) reconstruction, Mann only archived the spliced reconstruction segments in which, at each time step, the results of a later step were printed over results from earlier steps. Without the residual series no one could compute the unreported r2 scores.
What Was Being Hidden
In late 2003, only a few months after the “dirty laundry” email, we asked Mann to provide the residual series for the AD1400 step of his reconstruction. He refused. We filed a Materials Complaint to Nature, which had published the 1998 study, under their disclosure policies for either the residual series or the reconstruction steps. To their shame and discredit, Nature refused. We also requested the US National Science Foundation to require Mann to provide this data. To their discredit, they also refused.
Despite disinformation to the contrary, the results of Mann’s individual steps remain unarchived to this day.
We discovered the reason why Mann was so adamant about withholding his “dirty laundry” in 2004 – long before Climategate. By early 2004, despite many obstacles, we had been able to replicate Mann’s peculiar and poorly documented methodology well enough to calculate the residual series (and verification statistics) for the AD1400 step.
We discovered, to our considerable surprise, that the verification r2 statistic for the AD1400 step was disastrously low (0.018). The verification r2 is a commonplace statistic, which ought to be easily passed by any reconstruction purporting to have statistical “skill”. It is not a guarantee of model validity, but failure is more or less a guarantee of model invalidity. We reported our discoveries in two widely-discussed 2005 articles.2 At the time, we didn’t know for sure whether Mann had overlooked calculation of verification r2 values (implausible but possible) or whether he had calculated the values, discovered that they were disastrous and withheld them. Both alternatives were disquieting.
The dispute was prominently reported on in 2005, including a frontpage article in the Wall St Journal which attracted the attention of the US House Energy and Commerce Committee. They sent a set of questions to Mann including ones about source code and verification r2 statistics. These provoked vigorous protests from AAAS, AGU and other science institutions. Ralph Cicerone, then chair of the National Academy of Sciences, wrote to the House Energy and Commerce Committee offering their services, including, specifically, examination of the verification r2. Two studies were commissioned by congressional committees: the 2006 National Academy of Science and Wegman reports.
In partial response to the Committee questions, Mann archived some (but not all) source code for Mann et al 1998. While incomplete, it confirmed our surmise that Mann had calculated verification r2 statistics for each step of the signature NH reconstruction and had withheld them.
Subsequently, in 2006, Wahl and Ammann, both Mann allies and associates, did their own replication of the various steps: we were quickly able to reconcile their calculations to ours. They replicated the poor verification r2 for the AD1400 step and discovered that the score for the AD1600 step was even worse – perhaps the worst verification r2 in any scientific study ever published. Despite reconciling exactly to and confirming our results, their abstract misleadingly asserted that they had verified Mann’s results, when, in fact, they replicated ours – a point made at the time by Professor Wegman.
To this day, Mann has never archived the NH reconstructions for individual steps, the equivalent residual series (the “dirty laundry”) or even the verification r2 results.
No comments:
Post a Comment