The Age of Global Warming: A History (31 page)

Schneider’s probability claim underlines the importance of the pre-thermometer climate record in estimating the probability of whether mankind was causing the planet to warm which, he believed, computer models by themselves could not. It is also important in the history of the evolution of the scientific consensus on global warming. Before the scientific consensus claimed that the pre-thermometer record was unimportant when it had become discredited, it had claimed it was important.

Until the IPCC turned its attention to the pre-thermometer climate record, the widely accepted view was of considerable climatic variability, with a Medieval Warm Period starting around the turn of the first millennium, followed by falling temperatures and a Little Ice Age in the seventeenth century, followed by an erratic warming interspersed by reversions to colder conditions before a warming in the twentieth century. 

Hubert Lamb believed that the Medieval Warm Period, which was well documented in Europe, extended across the northern hemisphere. Lamb cited a Viking report of a cousin of Erik the Red swimming two miles across a fjord in Greenland, deducing that water temperatures were at least four degrees Celsius warmer than in the second half of the twentieth century.
[30]
Tree ring records from California indicated, in Lamb’s words, a sharp maximum of warmth, much as in Europe, between 1100 and 1300.
[31]

Warmer temperatures at more northerly latitudes coincided with a ‘moisture optimum’ in the Lake Chad basin, with a maximum occurrence of the pollen of monsoon zone flora. Population estimates for the Indian sub-continent also suggested sequences that roughly paralleled temperature trends in higher latitudes, falling temperatures there being mirrored in population declines in the sub-continent, Lamb thought.
[32]

The IPCC’s re-writing of the pre-thermometer climate record began in the Second Assessment Report. ‘Based on the incomplete evidence available, it is unlikely that global mean temperatures have varied by more than 1
o
C in a century’ since the end of the transition from the last ice age.
[33]
At that stage, revisionism was sporadic. The report also presented evidence that conflicted with this, supporting the views Lamb had expressed in his last book. Recent evidence from ice cores drilled through the Greenland ice sheet indicated that ‘changes in climate may often have been quite rapid and large, and not associated with any known external forcings’.
[34]
It also contained a warning that turned out to be prescient: ‘Because the high quality of much-needed long-time series of observations is often compromised, special care is required in interpretation,’ a warning the authors of the Third Assessment Report should have heeded.
[35]

In 1995, David Deming of the University of Oklahoma published a study inferring temperature records from boreholes which suggested North America had warmed somewhat since 1850.
[36]
As Deming later recalled, it gave him credibility within the community of scientists working on climate change. A major person working in the area emailed him: ‘We have to get rid of the Medieval Warm Period.’
[37]

Fred Pearce, the British environmental journalist and author, discovered in 1996 that Tim Barnett of San Diego’s Scripps Institution of Oceanography had teamed up with Phil Jones of University of East Anglia’s Climatic Research Unit to form a small group to mine paleo-climate records for signs of global warming and to summarise their research in the IPCC’s next assessment report. ‘What we hope is that the current patterns of temperature change prove distinctive, quite different from the patterns of natural variability in the past,’ Barnett told Pearce.
[38]

How could people be persuaded to believe that current temperatures were unnaturally high, when it was it widely believed that temperatures had been higher several centuries before the start of the Industrial Revolution? 

There was a way.

In 1912, fragments of the lower jaw of an orangutan, a human skull and a chimpanzee tooth were recovered from a quarry in the south-east of England. The bones, which had been stained with iron solution and acid, were taken to the British Museum. Man’s evolutionary missing link had been found. In the teeth of entrenched opposition from the paleontological establishment, an anatomist correctly identified Piltdown Man as a hoax, a finding which took a further thirty years to be definitively accepted.   

Tree rings from high altitude bristle cone pines would supply the raw evidence. In 1998, a paper written by a post-doctoral academic, Michael Mann, together with Raymond Bradley and Malcolm Hughes, based on Mann’s Ph.D. thesis, was published in
Nature
. It presented a new temperature reconstruction of the Northern Hemisphere going back six hundred years to 1400. The following year Mann (the sole expert who said that he could know the average global temperature to within 0.2
o
C) and his co-authors published a second paper, extending the series a further four hundred years to 1000. 

The picture it presented was of a stable climatic regime for the northern hemisphere, which lasted nine centuries, followed by unprecedented variability in the twentieth century, with the final decade averaging warmer temperatures than any time in the previous nine hundred and ninety years. The Medieval Warm Period had disappeared along with the Little Ice Age. The chart showed a long smooth line of low temperature variability trending downwards – in a 2007 Australian radio interview, Hughes claimed that the late nineteenth century was ‘one of the coolest, if not the coolest of the last several thousand years’ – followed by a sharp upward movement from around the beginning of the twentieth century.
[39]

The Hockey Stick gave the strong impression that the twentieth-century warming was anomalous. Equally the Medieval Warm Period and the Little Ice Age no longer needed to be explained or put within the envelope of natural variability because now they hadn’t happened. If Mann’s reconstruction was robust, it would help answer Lamb’s objection of the impossibility of defining a figure for natural variability and enable the IPCC to move decisively beyond the equivocal conclusion of the First Assessment Report – that the observed warming in the twentieth century had been within the bounds of natural variability.
[40]
 

There was one problem. It was too good to be true.

The series of tree rings which generated the hockey stick shape had been selected by a researcher, Donald Graybill, for an earlier study into the possible effects of carbon dioxide fertilisation. Graybill selected samples from bristlecone and closely related foxtail pines in the western United States for ‘cambial dieback’, where the bark had died around most of the circumference of the tree. Graybill and fellow researcher Sherwood Idso reported anomalously high twentieth-century growth for these compared to full bark trees unrelated to temperature data from nearby weather stations, a fact known to Mann and his co-authors.
[41]
Self-evidently they would make highly unreliable temperature proxies.

Their inclusion by Mann and his associates in their multi-proxy temperature reconstruction needed an additional step. To put measurements from different proxy series on a common basis to be statistically analysed, the data are standardised using a mean and standard deviation of the whole timespan of the series. At an early stage in the computational sequence, Mann used an algorithm derived from only the post-1901 portion of the data.
[42]
The effect was similar to having a Google search algorithm to give hockey stick-shaped series the highest ranking in the analysis.  

Did Mann know what he was doing?

He had graduated from Berkeley with a degree in applied mathematics and physics in 1989. Adopting this non-conventional approach to standardising proxy data would have required a conscious decision. To get the hockey stick from the data, Mann needed both the algorithm and Graybill’s tree ring data. 

Even if the algorithm had been selected erroneously, Mann had reason to know that producing a hockey stick depended on using the anomalous bristlecone ring series. Inside his directory of North American proxy data, Mann had a folder, BACKTO_1400-CENSORED, containing the North American data excluding all sixteen of Graybill’s series from the North American data. When the numbers from the CENSORED folder were run, the hockey stick disappeared.
[43]
If Mann hadn’t realised that his algorithm searched for hockey sticks, removing the Graybill series should have alerted him to the flaw in his method and meant that Mann’s public claims of robustness for the Hockey Stick were contradicted by what he’d found but buried in his CENSORED folder.

In 1998, Mann was nominated a lead author of the critical Chapter Two (‘Observed Climate Variability and Change’) of the Working Group I contribution to the Third Assessment Report, a signal achievement for someone only awarded his Ph.D. that year. The following year, he was appointed assistant professor at the University of Virginia. In 2000, the Hockey Stick got the highest possible political endorsement. In his final State of the Union message, President Clinton told Americans that the 1990s were the hottest decade of the entire millennium. Two years later,
Scientific American
named Mann one of fifty leading visionaries in science and technology.

Mann and his Hockey Stick didn’t have the entire field to himself. Competition came from Keith Briffa of the University of East Anglia’s Climatic Research Unit, who published a proxy temperature reconstruction within a couple of months of Mann’s first effort. Thanks to the Climategate emails released on the internet in November 2009, we now know how much tension this caused and the way the IPCC process resolved the issue in favour of Mann’s Hockey Stick. 

Briffa wrote a paper for
Science
comparing the two temperature reconstructions. After he’d sent a draft to Mann, in April 1999 Mann contacted the
Science
editor. ‘Better that nothing appear, than something unacceptable to us,’ he emailed, copying in Raymond Bradley, Mann’s co-author. Privately, Bradley immediately dissociated himself from Mann’s remarks. ‘As though we are the gatekeepers of all that is acceptable in the world of paleoclimatology seems amazingly arrogant,’ Bradley emailed Briffa.
[44]

After Briffa’s paper was published, Mann tried to patch things up. ‘Thanks for all the hard work,’ he emailed colleagues in May.
[45]
The sentiment didn’t go down well with Bradley. ‘Excuse me while I puke,’ Bradley emailed Briffa.
[46]

John Christy, who was also a Chapter Two lead author, recalled that the Hockey Stick featured prominently in IPCC meetings from 1999, telling the House Science, Space and Technology Committee in March 2011:

those not familiar with issues regarding reconstructions of this type (and even many who should have been) were truly enamoured by its depiction of temperature and sincerely wanted to believe it was the truth. Scepticism was virtually non-existent.
[47]

When the chapter lead authors met in Tanzania at the beginning of September, they were shown Mann’s Hockey Stick and Briffa’s reconstruction, which showed a sharp cooling trend after 1960.
[48]
The decline was more than a presentational problem. If tree ring growth declined when temperatures rose, what could be reliably inferred from them? And if the tree rings hadn’t responded to warming in the second half of the twentieth century, how could we know that they hadn’t done the same thing in response to possible medieval warmth?
[49]

In a follow-up email on 22
nd
September, IPCC coordinating author Chris Folland of the UK government’s Met Office said that a proxy diagram of temperature change was a ‘clear favourite’ for inclusion in the summary for policymakers. However, Briffa’s version ‘somewhat contradicts’ Mann’s and ‘dilutes the message rather significantly’, adding: ‘We want the truth. Mike [Mann] thinks it lies nearer his result (which seems in accord with what we know about worldwide mountain glaciers and, less clearly, suspect about solar variations).’
[50]

The same day, Briffa emailed acknowledging the pressure to present

a nice tidy story as regards ‘apparent unprecedented warming in a thousand years or more in the proxy data’ but in reality the situation is not quite so simple.
[51]

Briffa said he did not believe global mean annual temperatures had cooled progressively over thousands of years:

I contend that there is strong evidence for major changes in climate over the Holocene [the current geological epoch] (not Milankovich [climate changes induced by changes in the Earth’s orbit, tilt etc.]) that require explanation and that could represent part of the current or future variability of our climate.
[52]

The same day, Mann said he would be happy to add back Briffa’s temperature reconstruction. Including it risked giving ‘the sceptics’ a field day ‘casting doubt on our ability to understand the factors that influence these estimates and, thus, can undermine faith in the paleo-estimates’, unless there is a comment that ‘something else’ was responsible for the discrepancies.
[53]

Briffa hurriedly re-did his reconstruction.  He sent it to Mann at the beginning of October 1999, but the new version produced a larger post-1960 decline.
[54]

When the diagram was presented at the next lead author meeting in Auckland, New Zealand in February 2000, Briffa’s ‘disagreeable curve’ was not the same any more. It had been truncated around 1960; as Christy testified in 2011, ‘No one seemed to be alarmed (or in my case aware) that this had been done.’
[55]
And that is how the IPCC decided to present it in the Third Assessment Report, without noting the data deletion.
[56]

Star billing went to Mann’s Hockey Stick. A picture is worth a thousand words and the Hockey Stick appeared twice in the synthesis report, twice more in a diagram combining past and future temperature change and on the third page of the Working Group I Summary for Policy Makers.
[57]
The text placed the Hockey Stick into the context of progress since the Second Assessment Report. Working Group I’s contribution claimed that additional data from new studies of current and paleo-climates, together with ‘more rigorous evaluation of their quality’, had led to greater understanding of climate change.
[58]
Computer models suggested that it was ‘very unlikely’ (a term indicating a less than ten per cent chance) that the rise in observed temperatures could be explained by internal variability alone and the thousand-year temperature reconstructions indicated that this warming was unusual and unlikely (a ten to thirty-three per cent chance) to be entirely natural.
[59]
 

This opinion was translated to the front of the four-hundred-page synthesis report to answer the question: What is the evidence for, and causes of, changes in the Earth’s climate since the pre-industrial era? ‘The Earth’s climate system has demonstrably changed on both global and regional scales since the pre-industrial era, with some of these changes attributable to human activities,’ it stated in big, bold, blue letters.
[60]
  

The Hockey Stick had given the IPCC more than the single coin toss Schneider said was necessary to establish the probability of whether man-made global warming was real.