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 *