The Age of Global Warming: A History (30 page)

20

Turning Up The Heat

The scientists tell us the 1990s were the hottest decade of the entire millennium.

President Clinton, Address before a Joint Session of the Congress on the State of the Union, 27
th
January 2000

In holding scientific research and discovery in respect, as we should, we must also be alert to the equal and opposite danger that public policy could itself become the captive of a scientific-technological elite.

President Eisenhower, Farewell Address, 17
th
January 1961

Climatically, Kyoto was nigh on perfectly timed. According to one widely used global temperature series, February 1998 recorded the highest ever average global temperature in the instrumental record. 1998 was the warmest year so recorded.
[1]

The revised HadCRUT3 temperature series records a rise of slightly less than half a degree centigrade (0.496
o
C) from 1900 to 2000. Average observed temperature rose by 0.346
o
C from 1900 to 1944. Then, for almost a third of century, recorded average temperatures fell; the HadCRUT3 value for 1976 is actually lower than for 1900. It then rose by 0.526
o
C to 2000.

Put another way, over one hundred per cent of the recorded increase in average global temperature in the twentieth century occurred in its last quarter. 

So perhaps it’s not surprising that alarm about global warming rose through the 1990s.  Nonetheless, the multi-decadal fall in temperatures before 1976 required some kind of explanation, particularly to bolster the credibility of computer predictions of future temperature increase.

The IPCC sought to provide an answer to this conundrum in its Second Assessment Report. Doing so was important. If the climate models cited by the IPCC were unable to reproduce the broad outlines of the changes in global average temperatures, why should governments place much confidence in their ability to predict the future? Just because computer models could reproduce passable versions of the twentieth-century temperature record does not mean they will accurately foretell the future. But if they couldn’t reproduce the past, why should anyone have faith in them?

The Second Assessment Report claimed to have solved the puzzle. Including an assumption about the effects of sulphate aerosols enabled computer models to account for the thirty-year decline in observed temperatures from around 1944.
[2]
It also implied a higher value for the climate sensitivity of carbon dioxide. Computer simulations without aerosol forcing produced a best fit of a 1.5
o
C rise from a doubling of carbon dioxide in the atmosphere. Including sulphate aerosols resulted in the 4.5
o
C assumption providing the best fit.
[3]

How can we be sure that the assumption about sulphate aerosols wasn’t a case of adding epicycles? The Working Group I report recognised some shortcomings in knowledge. While there was a body of observations on aerosols, ‘The sensitivity of the optical properties to the size distribution of the particles, as well as to their chemical composition, made it difficult to relate the aerosol forcing in a simple manner.’
[4]
Additionally the global impact of aerosols on the thickness and extent of cloud cover was ‘recognised but could not be quantified’.
[5]

There was a larger issue the report was silent on – the validity of relying on computer models. From the Second Assessment Report onwards, the IPCC used the same language to describe computer model results as for empirical evidence derived from experiments on nature – ‘experiments with GCMs’ (general circulation models); ‘model experiments’; ‘recent multi-century model experiments’; ‘this hypothesis [dismissing the suppression of the greenhouse gas signature being masked by sulphate aerosols] is not supported by recent GCM experiments’.
[6]
Computer models thus attained an independent reality in the minds of climate scientists.

The Second Assessment Report noted that GCMs had to be calibrated by introducing systematic ‘flux adjustments’ to compensate for model errors.
[7]
In a post-Kyoto analysis, three distinguished MIT professors wrote that computer models were based on incomplete knowledge about the key factors that influence climate. ‘Today’s climate models cannot reproduce the succession of ice ages and warm periods over the last two hundred and fifty thousand years, let alone the smaller climatic fluctuations observed over the last century.’
[8]
Canadian mathematician Christopher Essex has described climate models as intricate pastiches to artfully simulate climatological phenomena in the absence of a comprehensive physical theory. Essex argued that climate models’ dependence on
ad hoc
constructions, known as parameterisations, meant that climate models could only represent the current climate regime where these
ad hoc
invariants hold.
[9]
Their values cannot be empirically validated for climate which had not happened.

A further challenge in trying to detect human impact on the climate was trying to define the envelope of natural variability, a problem that the climatologist Hubert Lamb believed to be practically insoluble. On the basis that natural variability plus anthropogenic forcing equals the instrumental record, the Second Assessment Report recounts naive attempts by climate scientists to calculate a value for natural variability by subtracting the assumed temperature changes in response to greenhouse gas forcing. To work out by how much extra carbon dioxide would warm the atmosphere, you need to know the amount of natural variability. To work out the amount of natural variability, you need to know by how much higher levels of carbon dioxide have been warming the atmosphere. As the IPCC noted, deriving an estimate for natural variability this way ‘depends critically’ on the accuracy of the models’ assumption of the climate sensitivity of carbon dioxide and other greenhouse gases.
[10]
The exercise risked toppling down the vortex of a circular argument.

In the absence of empirical data to fill critical assumptions, treating computer model results as
de facto
experimental evidence indicated a collective IPCC mindset that risked loss of contact with reality. After the IPCC Fourth Assessment Report, Bert Bolin wrote of new results from a ‘much richer set of experiments with global climate models’.
[11]
Commenting on a critical analysis by dissenting scientists, Bolin wrote: ‘The statement in their report that “computer simulations … can never be decisive as supporting evidence” is of course formally correct.’ Relying on them was justified because the IPCC’s conclusions were ‘still very plausible, being based on the wide variety of model experiments’.
[12]

This sort of thinking was something that shocked Professor Michael Kelly, a Cambridge University physicist and panellist in one of the 2010 Climategate investigations into the University of East Anglia’s Climatic Research Unit. ‘I take real exception to having simulation runs described as experiments (without at least the qualification of “computer” experiments),’ he emailed a colleague:

It does a disservice to centuries of real experimentation and allows simulations output to be considered as real data. This last is a very serious matter, as it can lead to the idea that ‘real data’ might be wrong simply because it disagrees with the models! That is turning centuries of science on its head.
[13]

The question the world most wanted the IPCC to answer was also the most difficult to give an unambiguous response: Were humans warming the planet and, if they were, by how much? The issue was covered in Chapter Eight of the Working Group I report. Statements on whether man-made global warming was actually happening are ‘inherently probabilistic in nature. They do not have simple “yes-or-no” answers’.
[14]

This should matter less for politicians than scientists, the IPCC suggested. A distinction should be drawn, the report argued, between ‘practically meaningful’ i.e., what policy should be based on, and what was ‘statistically unambiguous attribution’ to scientists.

While a scientist might require decades in order to reduce the risk of making an erroneous decision on climate change attribution to an acceptably low level (say one-five per cent), a policymaker must often make decisions without the benefit of waiting decades for near statistical certainty.
[15]

The report also sought to pre-empt counter-arguments based on scientific uncertainty.

The gradual emergence of an anthropogenic climate signal from the background noise of natural variability guarantees that any initial pronouncement that a change in the climate has been detected and attributed to specific causes will be questioned by some scientists.
[16]

The answer to the difficult question of when detection and attribution of human-induced climate change was likely to occur ‘must be subjective’.
[17]

Discussion on the Second Assessment Report’s top-line conclusion took a day and a half. Two of the Chapter Eight lead authors, Ben Santer and Tom Wigley, argued for a formulation saying that the balance of evidence pointed to an ‘appreciable human influence’ on global climate.
[18]
As chair, Bolin proposed ‘discernible’ instead of ‘appreciable’.

There were no objections to this proposal which better emphasised the uncertainty. Even though the precise meaning of the word ‘discernible’ was still somewhat unclear, to my mind it expressed considerable uncertainty as well as the common view that it was impossible to provide a more precise measure.
[19]

Interviewed during the Kyoto conference, Bolin went further. Global warming was not something ‘which you can prove’, going on to say: ‘You try to collect evidence and thereby a picture emerges.’
[20]

When published in June 1996, the Second Assessment Report ignited a row about the final text of Chapter Eight of the Working Group I report. Frederick Seitz, one of America’s foremost physicists and a former president of the National Academy of Sciences, wrote in the
Wall Street Journal
that in more than sixty years as a member of the American scientific community, he had never witnessed a ‘more disturbing corruption of the peer-review process’. After an IPCC meeting in Madrid the previous November, Seitz claimed that more than fifteen sections of Chapter Eight had been changed or deleted, including the sentence: ‘None of the studies above has shown clear evidence that we can attribute the observed [climate] changes to the specific cause of increases in greenhouse gases.’
[21]

Santer, as lead author of the Chapter, replied two weeks later in a letter co-signed with forty other scientists. They denied any procedural impropriety but admitted the draft had been changed at the behest of governments. ‘The changes made after the Madrid meeting were in response to written review comments received in October and November 1995 from governments,’ as well as from individual scientists and NGOs. Conceding that deletions had been made, the letter implied that they didn’t matter: ‘The basic content of these particular sentences has not been deleted.’
[22]
If so, why did the government reviewers insist on changing the text?

It is easy to see why the changes were politically necessary. The IPCC’s headline claim to have detected a discernible human influence on global climate would have been hard to square with the statement that no study had shown clear evidence that changes in the climate could be attributed to increases in greenhouse gases. A deputy assistant secretary at the US State Department, Day Mount, wrote to Sir John Houghton, chairman of the IPCC Working Group I, on 15
th
November: ‘It is essential … that chapter authors be prevailed upon to modify their text in an appropriate manner following discussion in Madrid.’
[23]

Santer and his fellow scientists were correct in suggesting that acceding to political pressure conformed to IPCC principles and procedures. In addition to nominating lead authors, governments effectively had the final say on what the IPCC published. Bolin had drafted the principles and procedures adopted by the IPCC in June 1993 for the Second Assessment Report. ‘Review is an essential part of the IPCC process,’ they stated. ‘Since the IPCC is an intergovernmental body, review of IPCC documents should involve both peer review by experts and review by governments.’
[24]
Advertised as summarising scientific knowledge, these procedures made IPCC reports politico-scientific documents in which scientific integrity took second place to political expediency in global warming’s hierarchy of needs.

This was made clear in a 1997 article co-authored by Stephen Schneider, who regularly acted as a spear-carrier for the global warming consensus. The Second Assessment Report, he wrote, ‘was fraught with political significance’ because it was published shortly before COP2 in Geneva at which the Clinton administration announced its support for binding emissions targets.
[25]
Rejecting the implication that rules had been broken, Schneider conceded that Seitz’s attack had ‘demonstrated that a hybrid science/policy organisation like the IPCC needs better, more explicit rules of procedure’.
[26]

Addressing the scientific issues underlying Seitz’s criticism, Schneider agreed with long-time critics such as Pat Michaels that direct observational evidence of global warming effects were not well-matched to climate predictions of one degree Celsius compared to the half degree observed in the twentieth century.
[27]
While critics put the rise down to natural variability, doing so without characterising the probability was, Schneider asserted, ‘a scientifically meaningless claim’. This is an odd position, as the ‘null hypothesis’, one where there is no statistical relationship between two phenomena, does not need to be explained. Odd, too, is Schneider’s criticism, given Bolin’s choice of ‘discernible’ to describe human influence on the global climate
because
of its imprecision. Schneider’s reasoning implies that the IPCC’s conclusion is also scientifically meaningless.

What, Schneider asked, is the probability that a half-degree warming trend was a natural accident? The answer could not be found in the thermometer record alone as it only stretched back a century. Doing so would be like trying to determine the probability of ‘heads’ by flipping a coin once. Climate scientists had to look at proxy records of climate change such as tree ring widths, deposits left by glaciers and ice cores. The proxy records, Schneider thought, suggested that a half-degree Celsius warming was not unprecedented, but neither was it common. ‘In my judgement,’ Schneider wrote, ‘this circumstantial evidence implies that a global surface warming of half a degree has about an eighty to ninety per cent likelihood of
not
being caused by the natural variability of the system.’
[28]

It was a bold claim. At a National Academy of Sciences hearing nine years later, a group of paleo-climatologists were asked by a member of the panel whether the science was such that they could determine the average century-scale temperature one thousand years ago within half a degree Celsius. Every presenter said ‘no’ — apart from one, who claimed to know the average century-scale temperature one millennia ago to within 0.2
0
C.
[29]