Smoking and climate change

Malcolm Turnbull, in a recent speech invented a parable about a foolhardy smoker to illustrate his claim that contesting the views of the “best scientists” on man-made climate-change is an “attack on science”. But does this parable make sense? Should we consider the the claimed “consensus” of science-policy bodies on dangerous global warming to be like the near unanimity of physicians on the effects of tobacco?

I have to say this is like ignoring the advice of your doctor to give up smoking and lose 10 kilos on the basis that somebody down the pub told you their uncle Ernie ate three pies a day and smoked a packet of cigarettes and lived to 95. Now that is how stupid it is and we have to get real about supporting and responsibly accepting the science. And if we want to challenge the science, do so on the basis of peer reviewed work of which I have to say, there isn’t a lot on the contrary side of the argument.

And this is actually — this war on science and on scientists which is being conducted is much worse than the case of person who ignores his doctor’s advice and follows the advice of his friend down the pub, drawing on the life experience of the fortunate Uncle Ernie.

Turnbull’s parable of the smoker is a rhetorical sleight of hand that doesn’t stand up to a second glance. He wants us to agree that casting doubt on anthropogenic causes of climate change is just like accepting a ridiculous assurance on the impact of smoking contrary to expert medical advice. But a moment’s thought shows that the two cases are not similar. Who, when a smoker justifies his habit by reference to a convenient anecdote, believes that this is meant as an honest or reflective rationale for smoking? We would recognise that the ridiculous justification offered in the Turnbull parable was rationalisation, not a reason: an excuse that the smoker used to deflect acceptance of the reality that their behaviour was ruled more by their addiction than by rational choice (I don’t’ mean that smokers are irrational).

Should we compare someone offering a rational argument on climate change with someone rationalising their chemical or emotional dependencies? Turnbull gives us no reason to do so. Of course, we could also imagine someone offering a trivial or ridiculous explanation for their dismissal of the scientific “consensus” on climate change (“someone down the pub told me…” etc.), but we would take this as simply a lack of serious interest in the subject, not a declaration of “war on science” as Turnbull, hyperbolically, claims.

Turnbull might have raised a more interesting question if he had used a more plausible example. Suppose, instead of relying on a someone’s uncle Ernie, Turnbull’s smoker had pointed out that almost 80% of smokers will not contract lung-cancer—their life-time risk is about 22%—and that he was prepared to accept those odds in return for the pleasure that he had from nicotine and from the act of smoking. This sounds more like a rational choice based, as it happens, on solid data. We could still call it a rationalisation of addiction because the argument makes selective use of well-attested evidence. But Turnbull would have had a hard time claiming that this statement, which justifies a choice on an unsound use of scientific evidence, amounted to an “attack on science.”

Could Turnbull have made a stronger case that anyone who accepts the unanimity of medical scientists on the noxious effects of smoking must also accept”consensus” on man-made climate change? Is the scientific case as strong for one as for the other? Malcolm Turnbull ducks and weaves around the question. At one point in his speech he says of the case for mitigation of global warming that “[t]he matter is simply one of risk-management”, which implies that there is no definitive evidence one way or the other. But in most of his speech he deplores any derogation from the “authority” of institutions and scientists who say the case for man-made climate change is a near certainty.

Although government now funds marketing campaigns intended to instil public dread of the consequences of smoking and climate-change—and public acceptance of taxes intended to affect personal choice in each domain—there is no comparison between the strength of the evidence offered for one and for the other. Indeed, the rationale for public action in each case reflects this difference. Because the evidence that smoking causes disease is strong, the social costs of smoking are readily calculated: about $21bn in Australia in 1999 split 65:35 between intangible costs (loss of life) and tangible costs to the productivity, health-care etc.[1] But the rationale for cutting emissions is only a precaution. The Garnaut Report recommends cuts to CO2 emissions on the basis of his estimate that the discounted costs will be smaller than the discounted benefits of avoiding a future catastrophe (his accounting is, nonetheless, implausible).

At first sight it seems, in principle, that we could be equally certain about the two propositions. They are similarly difficult to validate because it is impossible to perform “gold standard” randomised control, blinded experiments to test either. In the case of smoking, it would be un-ethical to randomly assign individuals to the “smoking” and “non-smoking” arms of a clinical trial, given the probability of harming those assigned to the “smoking” arm. In the case of climate-change we do not have a “laboratory bench” suitable for creating a experiment that can replicate the full complexity of planetary climate. So, in each case, we are forced to fall back on “natural” experiments: observations of past events, and predictions of future events that may be confirmed by un-biassed, prospective data collection.

But differences in the phenomena that are the subject of the “natural” experiments in each case mean that the results are not likely to be equally sound. The physiological mechanisms of the carcinogens in tobacco smoke are incompletely understood but they are likely to be linear. That is, we can expect there to be a direct proportionality between the scale of exposure and the scale of effect, sometimes called a “dose/effect” relationship. This means that cause and effect should be observed to be the same in repeated experiments. But the mechanisms of climate— also incompletely understood—are, fundamentally, non-linear. They are “chaotic” in the strict sense that the effects of any cause are sensitive to initial conditions surrounding that cause. There is no regular proportionality of cause and effect. To quote the Third Assessment report of the UNIPCC

These complex, chaotic, non-linear dynamics are an inherent aspect of the climate system. As the IPCC WGI Second Assessment Report (IPCC, 1996) (hereafter SAR) has previously noted, “future unexpected, large and rapid climate system changes (as have occurred in the past) are, by their nature, difficult to predict. This implies that future climate changes may also involve ‘surprises'”.

Although the climate system as a whole is chaotic, not every component is chaotic. The well-attested infra-red absorption effect of CO2 for example, shows a linear relationship of cause and effect, although the curve describes a logarithmic fall in the marginal effect as the volume of the gas increases. So we should expect to see some predictions related to man-made-climate theory verified or falsified by experiment. But because the system as a whole is chaotic, predictions about its overall direction must always be evaluated with much more caution than predictions in a linear framework such as those about the effects of smoking. We have to be especially careful in the case of chaotic systems to ensure that projections mesh with a very large number of observations and account for all significant observed trends (including reversals). Even so, the potential for “surprises” means our projections will have to be open to revision and (probably) not consolidated in a “consensus”.

Cycles affecting climate at different time-scales

Cycles affecting climate at different time-scales: source NOAA

It is possible to make confident predictions about, say, a smoker’s life-time risk of developing lung-cancer because there have been several large-scale long-baseline, prospective studies of cohorts of smokers and non-smokers carried out since the 1950s.[2] A male smoker in the United States has a cumulative annual risk of dying of lung-cancer before the age of 85 that is about 20 times the risk of a male who never smoked.[3] A fifty-year study of British doctors published in 2001 showed that about half died of smoking related disease and that deaths from lung-cancer were 15-times higher among those who smoked than among those who never smoked.[4]

But no evidence—much less prospectively collected evidence—directly correlates the activities of humans to changes in global temperature. Our direct observations of the climate are extremely brief in relation to the periods of known climate cycles such as the ice-ages of the holocene period (we are now, probably, near the end of the warm segment of the current “ice-age” cycle governed by the “orbital” and “Dansgaard/Oeschger” cycles in the image above). Inferences drawn from geological records of ice, sea-bed sediments, fossilised tree trunks etc suggest very large variations in the earth’s climate over long periods of time that are consistent with an inherently chaotic system.

The case for smoking as a cause of lung-cancer has been constructed methodically since the first statistical studies in the 1950s’s (Sir Richard Doll’s groundbreaking report did not postulate causation only correlation). The data is uncontaminated by time-dependent cycles and shows a strong correlation that has been confirmed frequently with a high degree of confidence by many different retrospective and prospective experiments among large groups of people. Cohort studies have evaluated even the least plausible alternatives to causation without success (see footnote 2 ). Physiological studies have also consistently added evidence of the carcinogenicity of tobacco-smoke (in many other tissues, as well as the lungs). Although much remains unexplained—such as why more than 70 percent of smokers will not die of lung-cancer—the underlying causal link between smoking and lung-cancer has been securely established by six decades of repeated experiment.

There is no comparable accumulation of unambiguous evidence of a human contribution to the warming of the atmosphere in the past 160 years (mostly in the past 40 years). Such data as we have on climate over that period is mostly from analog instruments whose accuracy (that is, error) is large compared to the observed change over that time: 0.7 degrees celsius +/- 0.2 degrees according to the UN IPCC. There is no apparent correlation of atmospheric temperature trends with CO2 concentrations in the atmosphere: CO2 levels rose at a steadily increasing rate in the past sixty years but temperatures rose, fell and stood flat for periods of a decade or more. We have been urged by a UN Committee to accept a theory of climate sensitivity to CO2 levels that depends on models whose results do not reproduce observed temperature changes without the help of “tuned” exogenous variables (in other words, fudge factors) such as the impact of aerosols. The crucial prediction of the models of tropospheric warming in low latitudes has not been confirmed. The models are unable to account for decadal cycles in temperature and pressure distributions that have a readily-observed impact on both warming and precipitation such as the El-Niño -Southern Ocillation (ENSO), the Pacific (PDO) and Atlantic Decadal Oscillations (ADO), etc. We have a growing body of evidence in “serious” science literature that the ancillary claims of the UN Committee intended to add to the plausibility of their modelling, including claims about sea-level rises, glacier retreats and sea-ice extent—are either false, not based on plausible evidence or exaggerated.

In summary, it is not only reasonable but prudent to be sceptical of an asserted “consensus” on the causes of global warming while accepting the unanimous opinion of physicians on the causal relations between smoking and cancer. Far from being an “attack” on science or scientists, a sceptical evaluation of the UN IPCC’s claims of a “highly likely” attribution of warming to human actions is warranted by a respect for science.


[1]Collins, D, and H Lapsley. “The social costs of smoking in Australia.” New South Wales public health bulletin (2004).[↑] [2]Some of the more recent studies are reviewed in: Cornfield, J, W Haenszel, E C Hammond, A M Lilienfeld, M B Shimkin, and E L Wynder. “Smoking and lung cancer: recent evidence and a discussion of some questions.” International journal of epidemiology 38, no. 5 (October 1, 2009): 1175–1191.[↑] [3]Thun, MJ, LM Hannan, and LL Adams-Campbell. “Lung cancer occurrence in never-smokers: an analysis of 13 cohorts and 22 cancer registry studies.” PLoS medicine (2008).[↑] [4]Doll, R. “Mortality in relation to smoking: 50 years’ observations on male British doctors.” BMJ 328, no. 7455 (June 26, 2004): 1519–0.[↑]

No Comments

Leave a Reply

Your email is never shared.Required fields are marked *