Prejudiced Authors Prejudiced Findings - 2

Similar scepticism was shown in numerous reviewers' comments for the second draft of chapter 9. The review editors appear to have ignored this skepticism. Chapter 9 makes scant mention of the differences of scientific opinion that exist. The IPCC’s procedural document contains an explicit requirement that significant differences should be described in an annex to the report. Yet, notwithstanding the absence of any scientific consensus on the magnitude of the human effect on the climate among suitably-qualified scientists, the only annex to chapter 9 deals with statistical methods.

According to the IPCC regulations, the review editors were required to present a written report to the IPCC Working Group Session (see quote above). Documents made available due to FOI requests by David Holland8 indicate that Karoly's report, in particular, consisted of a one-sentence letter with all the hallmarks of a standard form

http://www.climaeaudit.org/?p=2960

The addressees of Karoly’s letter reproduced above were the co-chairmen of the IPCC’s science “working group”. In fact, critical review comments were overridden or ignored entirely in the review process. As I have previously shown, the conclusions reached in chapter 9 were flawed and the evidence for them was insubstantial9.

Did the authors believe that the evidence was credible, or could other factors have influenced their thinking? Hegerl and Zwiers work with and develop climate models, the computer software from which climate predictions are made. The Hadley Centre and University of East Anglia, suppliers of 10 of the 53 authors of chapter 9, are likewise deeply involved with climate modeling, as are Allen of the University of Oxford and (probably) his three Oxford colleagues who also contributed to chapter 9. The US NCAR, supplier of 3 authors, also specializes in modeling. It is very likely that many other chapter 9 authors and the institutes to which they are affiliated are in the same position.

It would be unrealistic to expect that those who work with climate models would question the capability and accuracy of such models in the best of circumstances. The very heavy bias towards modelers among the authorship of chapter 9 must have largely prevented any serious questions about the competence of climate models (however sophisticated) to truly represent the future evolution of a complex, non-linear, chaotic object such as the climate. Indeed, it has been known since Lorenz (1963) that the initial state of the climate can never be known to a sufficient precision to allow reliable projections of its future evolution beyond a few days or weeks.

The IPCC's 2001 climate assessment played into the hands of the modelers when it said -

“... From a practical perspective, attribution of observed climate change to a given combination of human activity and natural influences requires another approach. This involves statistical analysis and the careful assessment of multiple lines of evidence to demonstrate, within a pre-specified margin of error, that the observed changes are:

 “unlikely to be due entirely to internal variability;

 “consistent with the estimated responses to the given combination of anthropogenic and natural forcing; and

 “not consistent with alternative, physically plausible explanations of recent climate change that exclude important elements of the given combination of forcings.”

 http://icecap.us/images/uploads/IPCC_evidence.doc

This statement would only be reasonable if climatologists had perfect knowledge of the initial state of the climate system, and of the processes by which it evolves. The first and most critical requirement is beyond any climate model, as the IPCC admitted in 2001 and again in 2007.

The words "model" and "models" appear a total of 628 times in the 70 pages of chapter 9, demonstrating how heavily the IPCC relies upon modeling for its findings. Chapter 9 does comment on the accuracy and completeness of climate models –

“Incomplete global data sets and remaining model uncertainties still restrict understanding of changes in extremes and attribution of changes to causes.” [p. 666]

“Detection of anthropogenic influence is not yet possible for all climate variables for a variety of reasons. Some variables respond less strongly to external forcing, or are less reliably modelled or observed.” [p. 669]

“Detection and attribution analyses show robust evidence for an anthropogenic influence on climate. However, some forcings are still omitted by many models and uncertainties remain in the treatment of those forcings that are included by the majority of models.” [p. 692]

However, these admitted deficiencies did not deter the authors of chapter 9 from presenting the output of climate models as if it was evidence of man-made warming.

Modeling is not evidence of anything, particularly when it is a chaotic object that is being modeled. Even then, it is only by the crudest of statistical prestidigitation that the authors of chapter 9 are able to reach their dubious conclusion that, in effect, the continuance in the past half-century of a warming trend that began a quarter of a millennium before humankind could have exercised any influence over the climate is caused by humankind.

Among the questionable devices deployed in chapter 9 are the following – 

First, chapter 9 asserts that historical annual mean temperatures are consistent with the climate models. However, the models have repeatedly been tuned ex post facto so as to reproduce phenomena (such as the cooling between 1940 and 1975) which they had not previously reproduced, so it is no surprise that – with the benefit of hindsight – the modelers have been able to force their models to replicate the observed climate. That process, however, cannot be relied upon to produce reliable projections for the future.

Secondly, the projections that are the output of the models are Delphic in their lack of specificity. When 58 simulations performed by 14 climate models deliver results often varying by more than 0.5 degrees Celsius (1 F: see, for instance, chapter 9, figure 9.5, p. 684), it is unsurprising that average global temperature falls within the range of outputs of the multitude of model simulations.

Thirdly, chapter 9 sets great store by “intercomparison studies” between the outputs of different climate models, implying that a “consensus” among climate models somehow constitutes scientific truth. William Briggs,10 a statistician, has shown that a consensus is to be expected between models because they share  similar mathematics, because knowledge is exchanged within the modeling fraternity, because models are deliberately tuned to match historical temperatures (Akasofu, 2008), and because any model that did not come close to reality would be discarded. The conditions precedent to “consensus” among models are thus in place even before the models start processing the data. 

Fourthly the authors of chapter 9 turn the scientific method on its head, stating or implying at many points that correlation implies causation. This is a well-known informal fallacy of logic, a variant of post hoc ergo propter hoc. It is generally true that lack of correlation necessarily entails lack of causation, but it is false that correlation necessarily entails causation. For instance the fact that atmospheric carbon dioxide concentration and global temperatures have both risen in the past half century does not necessarily imply either that the greater concentration of carbon dioxide caused the increase in temperature or that the increase in temperature caused the concentration of carbon dioxide to increase. Whenever two datasets appear to be correlated, causation of one by the other cannot be inferred until all possible causes extraneous to both have been considered and eliminated.

See http://wmbriggs.com/blog/2008/04/08/why-multiple-climate-model-agreement-is-not-that-exciting/

Chapter 9 makes several dubious claims that deserve a response. Here are just a few."

“The fact that no coupled model simulation so far has reproduced global temperature changes over the 20th century without anthropogenic forcing is strong evidence for the influence of humans on global climate. [pg 704] ... In addition, simulations forced with observed SST changes cannot fully explain the warming in the troposphere without increases in greenhouse gases, further strengthening the evidence that the warming does not originate from the ocean.” [p. 727]

An alternative and more plausible explanation is that the climate models are incomplete or inaccurate. Not one of the models predicted the stasis in global mean surface temperature since 1998, or the fall since late 2001. Some of the IPCC’s supporters have said that the current drop in temperatures is merely a temporary fluctuation consistent with natural variability of the climate, and that “global warming” will eventually resume, perhaps in 2015 (Keenlyside et al., 2008, vol. 453, 84-88). However, the IPCC and the fabricators of its models have always insisted that the role of natural variability in the climate is minuscule. If so, then they and their models are unable to explain the current fall in global temperature. 

“The ability of models to simulate many features of the observed temperature changes and variability at continental and sub-continental scales and the detection of anthropogenic effects on each of six continents provides stronger evidence of human influence on climate than was available to the [2001 report].” [p. 694] 

The output of a model might be acceptable evidence if it could be shown that the model was accurate in all respects and that there was no other plausible explanation. Throughout chapter 9 there is little recognition that the output of models should be taken as no more than indicative. What has been “detected” on each of six continents is a long-run trend of increasing temperature. However, the models are not capable of  distinguishing between natural and anthropogenic temperature changes, because the climate sensitivity to anthropogenic emissions of carbon dioxide is an input to the models, not an output from them (Akasofu, 2008).

The models are told at the outset to assume high climate sensitivity: without this assumption, it would not be possible to ascribe any temperature increase to humankind at all.

“A Bayesian analysis of seven climate models (Schnur and Hasselman, 2005) and Bayesian analyses of MMD 20C3M simulations (Min and Hense, 2006a,b) find decisive evidence for the influence of anthropogenic forcings.” [p. 689]

Bayesian analysis is a method of evaluating statistical probabilities. The analysis is not, as the IPCC says, an analysis of the models themselves but of their output. However, since it is the inputs to the models that determine the extent of the imagined human influence on climate, analyzing the outputs to determine the extent to which they demonstrate anthropogenic influences is meaningless and futile. All it indicates is that the models all make similar assumptions, whose validity Bayesian probabilities are entirely unable to verify. Besides, it is not appropriate to apply any form of probability theory (especially Bayesian probabilities) to the behavior of a chaotic object whose initial state and evolutionary processes are not known to a sufficient degree of precision.

“Schär et al. (2004) show that the central European heat wave of 2003 could also be consistent with model-predicted increases in temperature variability due to soil moisture and vegetation feedbacks. [p. 699] ... Based on a single study, and assuming a model-based estimate of temperature variability, past human influence may have more than doubled the risk of European mean summer temperatures as high as those recorded in 2003.” [p. 704]

However, chapter 3 of the IPCC’s 2007 climate assessment makes it explicit that the 2003 heat wave in Europe arose from a well-understood natural meteorological phenomenon – a stationary high-pressure cell spreading warm air from North Africa across Western Europe. The IPCC has rightly made it repeatedly plain that the attribution of individual extreme-weather events to anthropogenic “global warming” is not possible.

Nor is it appropriate to base conclusions about individual extreme-weather events on an exaggerated interpretation of a single academic study.

“Confidence [in the models’ assessment of internal climate variability] is further increased by systematic intercomparison of the ability of models to simulate the various modes of observed variability (Chapter 8), by comparisons between variability in observations and climate model data (Section 9.4) and by comparisons between proxy reconstructions and climate simulations of the last millennium (Chapter 6 and Section 9.3). ... Furthermore, because detection studies are statistical in nature, there is always some small possibility of spurious detection. The risk of such a possibility is reduced when corroborating lines of evidence provide a physically consistent view of the likely cause for the detected changes and render them less consistent with internal variability.” [p. 668]

Professor Richard Lindzen, who attended the sessions that led to the IPCC’s 2001 climate assessment, subsequently reported in testimony to the US Senate that officials of the IPCC at those sessions had circulated frequently among the groups of chapter authors, trying to persuade them that basing the entire case for climate alarm on computer modeling was an acceptable technique.

The above-quoted passage from chapter 9 is yet another instance of the IPCC’s undue and often naïve reliance upon modeling rather than upon theory and observation. As I have previously noted, intercomparison studies tend to embed errors systematically throughout all the models.

“The warming [from 1976 to 1998] occurred in both the ocean and the atmosphere and took place at a time when natural external forcing factors would likely have produced cooling.” [p. 665].

This statement characteristically ignores the fact that warming can be, and often is, caused by internal events, such as El Nino ocean oscillations like that which caused the global temperature peak in 1998. El Nino conditions have dominated the oceans and hence the climate since the mid-1970s, when the Pacific Decadal Oscillation moved into its warm phase. Contrary to the above quote, such conditions would likely have produced warming.

These unsatisfactory passages from chapter 9 demonstrate that the prejudice to be expected given the asymmetries and biases in the selection of chapter authors and reviewers is indeed present throughout the text. Chapter 9 is not science but politics elaborately dressed up as though it were science. The IPCC’s arguably apparent exclusion of scientists who had sufficient knowledge, impartiality, and integrity to prevent the numerous fundamental errors of science in chapter 9 has led to a statistically valueless attribution of the 1976-1998 “global warming” to humankind, when, on the evidence, it was merely the continuation of a natural warming trend that had set in some 300 years previously as solar activity recovered at the end of the Maunder Minimum.

That warming trend may now have ceased. During the 70 years from the 1930s to the 1990s, the Sun was more active, and for longer, than during almost any similar period over the past 11,400 years (Solanki et al., 2005).

However, the current 11-year solar cycle has got off to a slow start unprecedented since satellite observations began 30 years ago. Most days, no sunspots are visible on the solar surface, and the magnetic convection currents below the surface are inferred to have slowed considerably, presaging a global cooling that may begin in about a decade and continue for most of this century.

The IPCC tends to ignore such considerations. Its 2007 report chose an absurdly low estimate of the effect of solar variation on changes in global surface temperature, flying in the face of generations of data demonstrating a link between sunspots and terrestrial temperature. It was the astronomer Herschel, in 1801, who noticed an inverse correlation between the 11-year sunspot cycle and the price of grain as listed in a table in Adam Smith’s Inquiry into the Nature and Causes of the Wealth of Nations. Though correlation does not necessarily imply causation, there are good reasons for supposing that the Sun has a larger influence on the temperature of the solar system than the IPCC is willing to admit. “Global warming” has been observed or inferred not only on Earth but also on Mars, on Jupiter, on Neptune’s largest moon, and even on distant Pluto. Inferentially, the large, yellow object that gives the Solar system its name may have something to do with this simultaneous “global warming” on so many planets. Certainly, it would be rash to attribute extraterrestrial warming to anthropogenic enrichment of Earth’s atmosphere with carbon dioxide.

A recent image of the Sun, showing no sunspots at all.

For more than two years, sunspots have been intermittent. Though the first reversed-polarity, high-latitude sunspot of the new solar cycle was observed on 4 January 2008, there have been very few new-cycle sunspots since then, and there were none at all from mid-April to end-June 2008. This solar inactivity, following a period of intense solar activity almost without precedent in the 11,400 years since the end of the last Ice Age, presages cooler weather to come. To solar physicists, unlike the IPCC’s CO2-obsessed alarmists, the global cooling of the past seven years is unsurprising. The Sun is less busy, so the Earth begins to cool.

The hypothesis of damaging, man-made warming is a long way from being proven – and, given the recent trend in the peer-reviewed literature, is well on the way to being disproven. Recent cooling of the planet further suggests that man-made warming is at best too weak to be detected in the “noise” of natural internal variability.

To sum up, the IPCC is a single-interest organisation, whose charter assumes a widespread human influence on climate, rather than consideration of whether such influence may be negligible or missing altogether. Though the IPCC's principles also state that a wide range of views is to be sought when selecting lead authors and contributing authors, this rule has been honored more in the breach than in the observance.

More than two-thirds of all authors of chapter 9 of the IPCC’s 2007 climate-science assessment are part of a clique whose members have co-authored papers with each other and, we can surmise, very possibly at times acted as peer-reviewers for each other’s work. Of the 44 contributing authors, more than half have coauthored papers with the lead authors or coordinating lead authors of chapter 9. The IPCC appointed as review editor for chapter 9 a person who was not only a coordinating lead author for the corresponding chapter of the previous assessment report but had also authored 13 of the papers cited in chapter 9 and had co-authored papers with 10 authors of chapter 9 including both coordinating lead authors and three of the seven lead authors.

It is no surprise, therefore, that the majority of scientists who are skeptical of a human influence on climate significant enough to be damaging were unrepresented in the authorship of chapter 9. Most of the IPCC authors were climate modelers unwilling to admit that their models are neither accurate nor complete. Still less do they recognize or admit that modeling a chaotic object whose initial state and evolutionary processes are not known to a sufficient precision has a validation skill not significantly different from zero. In short, it cannot be done and has long been proven impossible. The modelers say that the “consensus” among their models is significant: but it is an artifact of ex-post-facto tuning to replicate historical temperatures, of repeated intercomparison studies, and of the authors’ shared belief in the unrealistically high estimate of climate sensitivity upon which all of the models rely.

It cannot be repeated too often that the supposed anthropogenic effect on global temperature is not an output from the models but an input to them. For this reason, they are wholly unable to shed any light whatsoever on the extent – if any – to which humankind may be altering the climate. The output of climate models, singly or as a group, is not evidence of anything. At best, it might be indicative, but it can never be conclusive.

Governments have naively and unwisely accepted the claims of a human influence on global temperatures made by a close-knit clique of a few dozen scientists, many of them climate modellers, as if such claims were representative of the opinion of the wider scientific community. On the evidence presented here, the IPCC’s selection of its chapter authors appears so prejudiced towards a predetermined outcome that it renders its scientific assessment of the climate suspect and its conclusions inappropriate for policy making.

John McLean is an IT professional in Melbourne, Australia, with over 5 years experience investigating and analyzing climate data and other climate-related issues. He can be reached at mcleanjoh@gmail.com