LL1-07 — Ch7 Halocarbons, the ozone layer and the precautionary principle#

Late lessons from early warnings: the precautionary principle 1896–2000 (EEA Environmental Issue Report No 22, 2001). Chapter 7, report pp. 76–83 (PDF pp. 76–83; printed and PDF page numbers coincide here).

How this was read. - I read the whole text extract page by page, through the final page marker (p. 83). - I then read all eight pages visually from the PDF, because the extract garbles Figures 7.1 and 7.2, splits footnotes across columns and loses the layout of Table 7.1. - I checked the report’s front matter (p. 12), the author note (p. 195) and the editors’ synthesis (pp. 169–187), but only for how they present this chapter. - External sources used for checking are listed at the end of the Limitations section.

Quotation policy. The report is marked “All rights reserved” (p. 2). These notes therefore summarise and paraphrase with page references. They use one short verbatim fragment and a number of two-to-five-word terms where the exact wording matters for the framing analysis. After the audits these include Farman’s hedges (“appear to be”, “appears to demand”, “can be seen perhaps”, “suggests”, “no known grounds”), because dropping them overstated his claims. Exact wording can be checked in the PDF at the pages cited.

Authors and standpoint#

• Sole author: Joe Farman (J. C. Farman, CBE, Polar Medal). The author note (p. 195) gives his position at the time as consultant to the European Ozone Research Coordinating Unit in Cambridge and member of the European Commission’s Science Panel on Stratospheric Ozone Research. It records that he worked for the British Antarctic Survey from 1956 to 1990, and lists UNEP and professional-society honours.
• He is a participant in the history he writes. He was lead author of the 1985 Nature paper reporting the Antarctic ozone losses (cited pp. 80, 83; Table 7.1). The editors flag this (p. 12): case-study authors are ‘not without strong views’, and Farman is their first example. They also say authors were asked to be brief and wrote pro bono, which explains the chapter’s compression.
• Evident stance.
• He is sceptical of the triumphalist account. He grants that Montreal is widely hailed as a success (p. 77) and calls 1987 a “psychological breakthrough” (p. 81). But he says the 1977–80 measures “appear to be” the only unequivocal precautionary action (p. 80), with one later “brief glimpse” at Beijing in 1999 (p. 81), and he reads Montreal as driven by events (p. 80).
• He is critical of consensus-paced negotiation, though his wording is hedged (“can be seen perhaps”, “in retrospect it seems”, p. 78). He is also critical of the industry-preferred HCFC/HFC transition. The latter is his one explicit first-person judgement: “In my view” (p. 81).
• He is critical of Du Pont’s evidential stance, but credits industry-funded research (p. 80) and a faster-than-expected industry response (p. 81).
• He is candid that the discovery surprised his own team (p. 82), and concedes that a 1965 risk assessment would have found nothing to worry about (p. 82). He is more restrained about foreseeability than the report’s editors are.
• Panels: none. The chapter has no panels or boxes. The only competing interpretation named is Richard Benedick’s; he was the chief US negotiator. Farman reports Benedick’s claim that Montreal applied the precautionary principle and offers “a simpler interpretation” against it (p. 80). Benedick’s work is not cited.
• Not all the chapter’s apparatus is Farman’s. Figure 7.1 and Table 7.1 are credited “Source: EEA” (pp. 76, 83), so they are probably editorial additions; Figure 7.2 is credited to DETR 1999 (p. 77). The Fig. 7.1 caption’s wording (including its scare quotes around ‘success’) and the Table 7.1 summaries should therefore not be attributed to Farman without qualification.
• Sources (p. 83).
• Mainly primary science: Houzeau 1857; Weigert 1907; Norrish and Neville 1934; Lovelock et al. 1973; Molina and Rowland 1974; Cicerone et al. 1974; Farman et al. 1985.
• WMO assessments 1985–99; UK DETR 1999; a 1987 US EPA Federal Register notice; one textbook (Goody 1954); one popular history (Cagin and Dray 1993); Du Pont’s 1975 newspaper statement (cited in the text, not the reference list); and an IISD web report on the Beijing meeting (fn 9, p. 81).
• Slaper et al. (1996), cited in the Fig. 7.1 caption, is missing from the reference list.
• No industry documents or social-science work on the negotiations are cited.

Section-by-section notes#

7.1 Overview (pp. 76–78)#

Scale and state of damage (p. 76). - Releases in the second half of the 20th century: about 23 Mt of CFCs, about 11 Mt of methyl chloroform, 2.5 Mt of carbon tetrachloride and 4 Mt of HCFC-22. The text mislabels HCFC-22 as “a hydrofluorocarbon”; footnote 7 on p. 81 draws the distinction correctly. - All of these live long enough to reach the stratosphere. Footnote 4 defines lifetimes as e-folding times. - By 2001, stratospheric chlorine was six to seven times its 1950 level. There was an Antarctic hole every September to December, large but irregular Arctic losses, and moderate losses at mid-latitudes in both hemispheres. - Causation is said to be established “beyond reasonable doubt” by five WMO assessments (1985–99). This is a legal-style standard of proof. Impacts are expected to include more skin cancer.

Figure 7.1 (p. 76; source EEA, caption citing Slaper et al., 1996). An explicitly schematic chart (“not to scale”), 1975–2100: - CFC production peaks around 1988–90 and ends around 2005. - Stratospheric ozone-depleting substances peak in the late 1990s and decline slowly. - Surface UV excess peaks around 2000 and fades by the 2050s. - A rising background skin-cancer line ends in a question mark. - Additional skin cancer from ozone depletion starts around 2010, peaks around 2050 and fades by the 2080s, reflecting a 30–40 year latency.

The caption lists the complications: other depleting gases; variation by latitude, season and weather; wavelength and cloud effects; melanoma versus other cancers; behaviour; cataracts and immune effects. It ends by saying the figure shows the ‘success’ (in quotation marks) of stopping CFC production, which averted much more skin cancer than is now expected. So the caption makes a positive claim about the Protocol’s benefit; the quotation marks qualify it rather than dismiss it. The figure and caption are credited to the EEA, not Farman (see above). Reading note: this is the chapter’s clearest picture of chained time lags, with harm peaking about 60 years after production peaks.

“End of the CFC episode” (p. 77). - Consumption (production plus imports minus exports) and most production of CFCs, halons and methyl chloroform had ceased in developed countries. - The exceptions were feedstocks, “so-called” essential uses (footnote 5: medical inhalers, by then largely switched to HFCs or PFCs, and military halon fire protection), and supply for developing countries, whose freeze began in 1999 with cessation due in 2010. - The Protocol was signed in 1987 and in force from 1989, and has been hailed as a major success.

Persistence and banks (p. 77). - Methyl chloroform has a lifetime of about 5 years, so about 2% remains after 20 years. - CFC-12 has a lifetime of about 100 years, so it will be at least 37% of today’s level in 2100. - That is a floor, because equipment and foam banks (791 kt in 1995) keep leaking.

An unnamed halon (p. 77). The text then switches substance without naming it. The passage covers: - production peaking in 1988 and ending in developed countries at the end of 1993; - some production in China, India and Korea, due to end in 2002; - a developed-country bank of about 70 kt in 1995, run down by about 4% a year; - a lifetime of about 65 years, with concentrations rising until about 2020 unless the Protocol requires banked halons to be destroyed.

The gap is in the printed PDF and its text layer, not just the extract (rechecked in the second audit). It is almost certainly halon-1301. This is my inference: of the major halons, only halon-1301 has a lifetime near 65 years in assessments of the period, while halon-1211’s is roughly a decade or so (ext.). The developed-country end-1993 phase-out date is consistent.

Figure 7.2 (p. 77; source DETR 1999). Tropospheric chlorine and bromine, 1940–2100. Combined equivalent chlorine crosses the dashed 2.5 ppbv line in the late 1970s (about 1977 on the chart; chlorine alone crosses it about 1981). It peaks near 4.5 ppb in the mid-1990s and returns to 2.5 in the late 2050s. Chlorine alone peaks near 3.6 ppb. Bromine peaks around 2000.

Equivalent chlorine and adequacy (p. 78). - Equivalent chlorine = Cl + 58 × Br, since bromine is about 58 times more destructive per molecule. - By convention, 2.5 ppbv (first reached in the late 1970s) is the significant level, though opinion differs on when damage began. - Under the 1997 amendments, the return to 2.5 is expected in 2050–60. - Key claim: under the original Protocol, London (1990) and Copenhagen (1992), there was no guarantee of ever returning to that level. Vienna (1995) first achieved a projected return, and 1997 brought it “a few years nearer”. The chapter gives no source for this sequence of projections; it rests on Farman’s expert authority and the WMO assessments generally.

Diagnosis of the negotiations (p. 78). Farman’s wording is hedged. The failure to act more decisively “can be seen perhaps” as the Parties putting ‘consensus before effectiveness’; “in retrospect it seems” the negotiators consistently assumed agreement would be easier at the next meeting. Footnote 6 supports this: - the last-resort voting rule (Art. 2 para 9(c)) had never been used; - the clause allowing stricter measures (Art. 2 para 11) had been formally invoked only once (by “the European Union”, March 1991). - Farman’s own inference from this: it suggests that, on the whole, the negotiators were “quite content” to proceed at a pace compatible with consensus.

Check: the voting text Farman quotes is the 1987 original, already amended in 1990 (see Limitations). The section closes by asking whether the CFC episode should, or could, have been avoided. It does not answer directly: it offers three historical sketches that may help to “put these questions in perspective” (p. 78).

7.2 Early history, c. 1900 (pp. 78–79)#

Chemistry. - Chlorinated hydrocarbons studied in the 19th century included many compounds with no significant sources outside the laboratory, as with most ozone-depleting substances (p. 78). Reading note: novelty as a marker. - Carbon tetrachloride was already an industrial solvent and extinguishant; methyl chloroform was known but unused. - Swarts (Belgium) had made CFC-11 and CFC-12.

Ozone as a product (p. 78). Around 1900 ozone was an important industrial chemical, in an episode Farman says is “now largely ignored in textbooks”. It was the most powerful oxidant known, left no objectionable residue, and was easy, if somewhat expensive, to make. It was used for water sterilisation, air in the London Underground, bleaching and deodorising in textiles, paint and perfume manufacture, and food preservation. Reading note: society used ozone as a product long before it understood ozone as a shield.

Atmospheric knowledge (pp. 78–79). - Ground ozone had been measured since 1857 (Rouen, about 10 ppbv), and higher concentrations aloft were inferred from the solar spectrum. - Meteorologists’ comfortable belief in a simple thermal structure was upset by Teisserenc de Bort’s discovery of the stratosphere, dated 1901 in the text. - Routine stratospheric soundings began in 1933 but were held back by recession and war until the 1950s.

Reading note: the first recurring pattern is confident expert consensus about a simple system overturned by observation.

7.3 The 1930s: the CFC industry is born (p. 79)#

Industry. - Cheap chlorine, a by-product of making caustic soda by electrolysis of salt, had displaced ozone. - Rotheim (spelled “Eric”) applied for a patent on the aerosol-container principle in 1926. It was little used until Second World War insecticide sprays; mass production began in the US in 1947 and in Germany in 1953. - In 1929 Midgley (spelled “Midgely”) at General Motors recommended CFC-12 and CFC-11 as efficient, non-toxic and non-flammable refrigerants. Du Pont began production of CFC-12 in 1930 and CFC-11 in 1934. - Reading note: the selection criteria were performance plus safety in handling. Environmental fate did not feature.

The science at launch. - The ozone layer had been discovered and, “considering the resources available”, “quite comprehensively investigated”. Dobson and his Oxford collaborators, following Fabry and Buisson’s pioneering measurement, had stations from the Arctic Circle to New Zealand. Two very different vertical-profile methods (zenith-sky twilight measurements; the Regeners’ balloon spectroscope to 31 km) agreed remarkably well. - Chapman’s 1930 photochemical theory fitted the upper layer but failed completely on variation with latitude and season. Realistic air motions entered models only more than 50 years later, and chemistry–dynamics uncertainty persisted into 2001. - Reading note: when CFCs went on sale, the relevant science was in no position to assess their fate in the stratosphere. The isolated laboratory hints of 1907 and 1934 (p. 82) were not connected to CFCs. This underpins the closing claim that technology outstrips risk science (p. 83).

7.4 The 1970s: “the seeds of doubt” (pp. 79–80)#

The warnings (p. 79). - Ozone became an international issue in 1970, over emissions from planned supersonic transport fleets. Farman says the fears were “heeded”: the US abandoned its SST, and Concorde plus the Tu-144 never exceeded 40 aircraft. This is a simplification of the US decision (see Limitations). - CFC use had grown enormously in the 1960s. Lovelock et al. (1973), using a new trace detector, found CFCs worldwide, with almost none removed. - In 1974, Molina and Rowland and, separately, Cicerone et al. proposed the mechanism: stable CFCs reach the stratosphere, photolysis frees chlorine, and a chain reaction destroys ozone. More than a decade of vigorous debate followed.

The first responses (pp. 79–80). - 1977: UNEP’s research-oriented World Plan of Action and its Coordinating Committee on the Ozone Layer, with NGOs allowed to attend. Farman calls this “the first significant move” (p. 79). - A US public campaign, primed by the SST debate, produced 1977 rules banning CFC aerosol propellants. Canada, Norway and Sweden followed. - European environmental groups were concentrating on acid rain, and there was little public pressure over CFCs. Even so, Council Decision 80/372 (1980) aimed to freeze CFC-11 and CFC-12 production capacity and to cut aerosol use by at least 30% from 1976 levels by the end of 1981 (p. 80).

Why the early measures’ effect is hard to see (p. 80). Farman says their effect on releases “is not easily seen”: - Releases had already fallen sharply a few years earlier, largely because of the downturn in world trade after the oil crisis. - The cut in aerosols was offset by growth in foam blowing. (Table 7.1 on p. 83 says refrigerators instead, a small inconsistency.) - The EC capacity freeze was a ‘token gesture’, since plants were running below capacity.

The precautionary verdict (p. 80). - Farman nonetheless says these measures “appear to be” the ‘first, and last’ unequivocal use of precaution in the story, and the only action taken before the evidence became compelling. “These measures” covers both the US/Nordic bans and the EC decision he has just called a token gesture. - He contrasts the 1977 US Clean Air Act’s ‘reasonable expectation’ standard (cited via US EPA 1987) with industry’s ‘wait and see’. - Du Pont, which Farman calls the world’s first and largest CFC producer, set out that attitude in a full-page New York Times statement of 30 June 1975. It said it was prepared to stop production of the “offending compounds” if ‘reputable evidence’ showed that some fluorocarbons caused a health hazard through ozone depletion. In Farman’s account, Du Pont denied such evidence existed until 1986. - In fairness, he notes, industry funded research substantially throughout, via the Chemical Manufacturers Association.

The early-1980s lull and the Vienna Convention (p. 80). - The problem seemed settled: models predicted small losses and observations showed no significant trend. - UNEP negotiations from 1981 split the US (controls on particular uses) from Europe (a cap on capacity). - The Vienna Convention (March 1985) contained only commitments to research, monitoring and information exchange, with controls “if and when justified”. - By this time (1985) the US was not prepared to act unilaterally. It claimed that its earlier action had cut its share of world CFC-11 and CFC-12 production from 46% (1974) to 28% (1985). Farman reports the causal attribution as a US claim, not as his own finding. - UNEP was mandated to negotiate a protocol for 1987.

7.5 The Montreal Protocol and the ozone hole (pp. 80–81)#

Two readings of the signing (p. 80). The Protocol was signed on 16 September 1987 after a contentious final week. - Benedick: later called it an application of the precautionary principle. - Farman: the negotiators were ‘overtaken by events’. His evidence: - The May 1985 Antarctic report showed losses far beyond any prediction, confirmed by NASA in October 1985. The Washington Post coined ‘ozone hole’. - Du Pont, reminded of its 1975 pledge, told customers in September 1986 that it accepted the need for some controls (Cagin and Dray, p. 308). - In 1986, NOZE at McMurdo found evidence that the loss was chemically driven. - In September 1987, results were expected from NOZE II and the airborne experiment flying from Punta Arenas. - Farman’s conclusion: the timing makes sense only as a ‘pre-emptive move’, “astutely designed” to protect the negotiators’ credibility and give industry time for orderly reorganisation. “Astutely” makes this a qualified compliment as well as a debunking. This is inference from timing; no documents are cited.

Assessment of 1987 and London 1990 (p. 81). - Various commentators called the 1987 Protocol a success, a compromise, a muddle and a failure. Farman calls it a ‘psychological breakthrough’, but its terms were limited to what seemed practicable, its goal was unclear, and releases rose rapidly before the next meeting. - In force from 1 January 1989, with review starting at once. By then there was scientific consensus, active NGO campaigns, and industry responding much faster than expected. - London (1990) was stronger but disappointing against the Parties’ own preparatory statements. It brought two important institutional changes: full reviews every two years instead of four, and the Article 10 Multilateral Fund (MFMP) for developing countries.

The substitution critique, the chapter’s central policy argument (p. 81). - The pathway: negotiators prioritised quick replacement of CFCs with the substitutes industry preferred. Footnote 7: HCFCs deplete ozone, though less; HFCs do not, but are potent greenhouse gases. - The interest: about 75% of CFC output was held by 13 company groups. They were willing to close old plants given reasonable time to profit from HCFC and HFC investment. - The deal: negotiators readily agreed. The transitional substances got guidelines, not controls, and an open-ended future, because there was no consensus on a phase-out date (footnote 8: proposals ranged from 2010 to 2040). - Farman’s judgement: ‘deeply flawed’. - Surveys showed large avoidable releases from poor practice, so less replacement was needed than consumption implied. - Policy should have set prudent long-term goals and actively fostered halocarbon-free, energy-efficient technology, which would have served the ozone layer, the climate and developing countries’ costs at once. - Fund money went mainly to replacing CFCs with HCFCs, and phasing out HCFCs would now need more. - Radical technological change should have been stimulated from the start.

Beijing, December 1999 (p. 81). - 129 governments attended. - Before the meeting, the MFMP approved USD 150 million to close China’s CFC production over 10 years (China was then the largest producer and consumer of CFCs and halons) and USD 82 million for India (the second-largest producer). - The Parties agreed a further USD 440 million; the fund had disbursed more than USD 1 billion since 1991. - New HCFC controls, including a trade ban with countries that had not ratified the 1992 amendment; bromochloromethane production to end by 2002; and a declaration on illegal trade. - The “one brief glimpse” of precaution: the assessment panels were asked to propose ways to prevent the development and marketing of new ozone-depleting chemicals. - Unresolved: destroying halon banks, an earlier methyl bromide phase-out, and consumption in Russia and other economies in transition.

7.6 Late lessons (pp. 82–83)#

Discovery (p. 82). - The 1985 finding came from systematic long-term measurements begun for exploration, and surprised everyone, including the authors. - Theory expected the first effects high in the tropical stratosphere (30–50 km), with slow polar change. - Two other groups were close: Japanese scientists had reported anomalous Antarctic profiles without identifying a trend, and NASA was re-examining very low satellite values that its software had flagged as ‘suspect’. - Why the threads were not joined up: no Internet, stretched institutions and the workload of the first WMO assessment. - Long-term monitoring funding is a serious problem, but open-ended funding at the level needed is not practicable.

Growth and “wonder chemicals” (p. 82). - Cumulative releases from 1930 to 1948 were 25 kt (CFC-12) and 5 kt (CFC-11). By 1970 annual releases were 300 kt and 207 kt. - Recession and war had held back the industry’s early growth. - Aerosols persuaded people to buy CFCs and throw them away; industry’s dream of ‘wonder chemicals’ seemed to be coming true. - That, Farman says, was when it “might reasonably have been asked” whether such development was sustainable.

The 1965 counterfactual (p. 82). - There “can be little doubt” that a conventional assessment would have found no known grounds for concern. CFCs were chemically very inert, non-flammable, of very low toxicity and useful (heat insulators, solvents). - The assessment might have noted that nobody knew what happens to CFCs once released to the atmosphere. But it would “no doubt” have added that they had been released for more than 30 years with no apparent harm. Reading note: the gap in knowledge would have been acknowledged, then discounted against the track record. - It would have recommended further research, which in hindsight should have covered UV and IR spectra, atmospheric concentrations and breakdown products. - Laboratory hints existed (Weigert 1907; Norrish and Neville 1934). But they were relevant only if someone had recognised that atomic chlorine and fluorine might be released from CFCs in the upper atmosphere. Footnote 10: the papers were cited in a 1977 review of stratospheric chemistry, but nothing suggests they influenced workers in the 1960s and early 1970s. Even then, Farman says, they “would surely have been dismissed” because there would be at least 10,000 times more ozone than CFCs at the relevant altitudes. - It is “just possible” that such an assessment might have sped up the research needed for a plausible case. But a plausible case brought only limited action anyway. Serious negotiation waited for severe depletion and strong evidence, and the protracted, frequently amended schedule “suggests” that precaution was not uppermost in policy-makers’ minds. - Check: by Farman’s own account (p. 80), UNEP convention talks began in 1981 and the Parties authorised protocol negotiations in March 1985, two months before the Antarctic paper. The claim therefore rests on the word “serious”. It is a judgement about the quality of the negotiations, not a claim that none took place before the hole. - Reading note: the chapter never answers its own opening questions directly (should, and could, the CFC episode have been avoided? p. 78). The 1965 counterfactual implies that conventional assessment could not easily have avoided it. The ‘no mandate for global experiments’ principle implies that it should have been approached differently. That second answer is normative, not evidential.

“Global experiments” (pp. 82–83). - The US decision in 2001 not to ratify Kyoto disappointed those who hoped Montreal would ease further agreements. - The lesson should have been that neither governments nor multinational companies have a mandate for global experiments, even ‘business as usual’. - The CFC issue is a “stark warning”. Practices that appear to be reasonable when introduced, amid huge gaps in understanding, may later be seen, as understanding improves, to be leading to a major global problem that can be neither avoided nor rapidly alleviated.

The paradox and the close (p. 83). - Short-term safety appears to demand that synthetic chemicals in everyday use be non-reactive. It took a long while to realise that this means they will be extremely persistent. Farman calls this a “deep-seated paradox”. - Science cannot foresee all hazards, and technology often outstrips the science needed to assess it. - “If there is to be a solution”, policy-makers must learn to recognise much sooner “when ignorance has been replaced by understanding, however rudimentary” (p. 83; the only full-clause verbatim quotation in these notes). The conditional framing matters: Farman offers this as a necessary condition, not a confident remedy.

Table 7.1 “Halocarbons: early warnings and actions” (p. 83; source EEA)#

The editors’ table lists 13 dated entries: 1907, 1934, 1973, 1974, 1977 (the US ban “based on ‘reasonable expectation’”), 1977 (UNEP plan), 1980, 1985 (Vienna), 1985 (Farman et al.; “Antartica” [sic]), 1987, 1990s (finance), 1997 (return of chlorine by 2050–60) and 1999 (Beijing Declaration).

Omissions: Cicerone et al.; Du Pont’s 1975 and 1986 positions; NASA’s confirmation; London 1990, Copenhagen 1992 and Vienna 1995; and the cessation dates. It is a list of events, not of contests.

Overstatement in the 1987 entry. The table says the Protocol was signed “with phasing out” of ozone-depleting substances. The original 1987 text set freezes and staged reductions (for CFCs, a 50% cut by 1998–99), not a full phase-out; phase-out came with the 1990 London amendments (ext.). The chapter’s own text implies as much: under the original Protocol there was no guarantee that chlorine would ever return to 2.5 ppbv (p. 78).

Case timeline#

External facts are marked “ext.”; all other entries are from the chapter.

Lags from the first credible warning (1974):

From the observed Antarctic loss (1985) to international controls took only about 2 years. That contrast is one of the chapter’s central points: a plausible mechanism brought “only limited action”, while serious negotiation began only once severe depletion had occurred and strong evidence linked it to CFCs (p. 82). It should not be overstated. Farman does not say observation produced decisive action. He says the early negotiations failed “to act more decisively” (p. 78); the 1987 terms were limited to what was thought practicable (p. 81); and the schedule stayed protracted (p. 82). From commercial launch (1930) to the first credible warning took 44 years. In that period the ozone layer was “quite comprehensively investigated” for the resources available (p. 79), but the pieces needed to assess CFCs were missing or unconnected: trace-gas measurement, stratospheric chlorine chemistry and realistic dynamics (pp. 79, 82).

What was known when: - 1930: the ozone layer and the stratosphere were known; the fate of CFCs was unknown (pp. 78–79; p. 82 says it was still unknown in 1965). - 1973: persistence and global spread were measured (p. 79). - 1974: a mechanism had been proposed (p. 79). - 1985: damage far beyond the models was observed (p. 80). - 1986–87: the chemical cause was increasingly supported (p. 80). - 1989: scientific consensus on the main issues (p. 81). - 2001: causation established “beyond reasonable doubt” (p. 76).

Harms and costs as stated: - Health: skin cancer, cataracts and immune suppression, with 30–40 year latency and excess peaking around 2050 (Fig. 7.1, p. 76). - Environment: ozone losses across latitudes (p. 76). - Finance: more than USD 1 billion disbursed by the MFMP, with more needed for the HCFC phase-out (p. 81). - Not quantified: health burden, damage costs, control costs.

The authors’ own lessons and conclusions#

(Sole author: Farman.)

Derived from the chapter’s evidence#

1. Precaution was barely used. The 1977–80 measures “appear to be” the only unequivocal case, and were weak in effect (p. 80). Montreal was driven by events (p. 80). Beijing 1999 offered one “glimpse” (p. 81). The drawn-out schedule “suggests” precaution was not uppermost (p. 82).
2. A plausible case brought only limited action. Serious negotiation waited for severe damage and strong evidence (p. 82).
3. Consensus trumped effectiveness (hedged: “can be seen perhaps”, “in retrospect it seems”). There was no guaranteed recovery until 1995 (p. 78; fn 6).
4. Ending production does not end the damage, because of lifetimes and banks (p. 77; Figs 7.1–7.2).
5. Discovery came from long-term curiosity-driven monitoring and surprised everyone. Theory pointed elsewhere, and scattered anomalies were not joined up because of stretched resources, weak collaboration and assessment workload (p. 82).
6. Conventional risk assessment in 1965 would have found no known grounds for concern. The unknown atmospheric fate of CFCs would have been outweighed by 30 years without apparent harm. Early hints would surely have been dismissed on relative abundance, and even a better assessment might only just possibly have sped up research, not action (p. 82).
7. The inertness–persistence paradox: short-term safety “appears to demand” non-reactive chemicals, which will be extremely persistent (p. 83).
8. Technology outstrips the science needed to assess it (pp. 79, 83).

Recommendations and advocacy (normative)#

• Recognise much sooner when ignorance has become rudimentary understanding (p. 83). This is the main recommendation. That policy-makers should then act is implied rather than stated.
• Neither governments nor multinational companies have a mandate for global experiments, even ones that consist solely of ‘business as usual’. Farman frames it as a lesson that “should have been” learned. He applies it to climate policy and US non-ratification of Kyoto (p. 82).
• Phase-outs should aim at halocarbon-free, energy-efficient technology rather than chemical-for-chemical substitution. They should set long-term goals, cut avoidable releases, and (the notes’ gloss) use development finance to leapfrog rather than to fund CFC-to-HCFC substitution (p. 81).
• Destroy halon banks, bring forward the methyl bromide phase-out, and tackle consumption in economies in transition. These are implied by listing them as unresolved (pp. 77, 81).
• Monitoring funding is a serious problem. But open-ended funding is “not practicable”, and no remedy is proposed (p. 82).
• Implicitly, decision rules should be more decisive than consensus. Implied by the critique of the unused voting and stricter-measures clauses (p. 78, fn 6). No reform is specified.

Assessment. Items 1–8 are grounded in the chapter’s own (compressed) evidence. The substitution recommendation rests on uncited “technical surveys” (p. 81) and an untested counterfactual. The “no mandate” claim is a normative principle.

Mechanisms and dynamics#

1. Knowledge production and surprise#

• Ignorance, not uncertainty, governed 1930–1973. At launch, the relevant science could not frame the question: soundings were rare and ozone theory incomplete (pp. 78–79). The tools that exposed the problem, a new trace-gas detector and a chlorine-chemistry mechanism, arrived only in 1973–74 (p. 79). The editors draw the ignorance point explicitly (p. 169).
• Expectations steered attention. The accepted view put the first signal in the tropical upper stratosphere (p. 82). In the early 1980s, models and observations agreed that little was happening (p. 80). That agreement reassured, but about the wrong place and the wrong mechanism. The chapter never explains why the models missed polar loss (the chemistry on polar stratospheric clouds is not mentioned).
• Screening encoded priors. Satellite software flagged very low values as ‘suspect’ (p. 82). Automated plausibility checks embody their designers’ sense of what is possible.
• Systematic long-term records. The decisive evidence came from a series begun for exploration (p. 82). Farman stresses its systematic continuity; the editors stress serendipity (pp. 172–173).
• Institutional friction. British, Japanese and NASA groups each held part of the anomaly. Farman cites the pre-Internet era, an absence of effective collaboration (“regrettable” but “not uncommon” in institutions with stretched resources) and the workload of the first WMO assessment (p. 82). Notes’ gloss: consolidating existing knowledge competed with noticing new knowledge. Farman only says the WMO report “should also be noted”.

2. Burden and standard of proof#

• Two thresholds in direct conflict (p. 80): the statutory ‘reasonable expectation’ of harm, and the producer’s own ‘reputable evidence’ test. The pledge was conditional on evidence the producer itself judged. By Farman’s account Du Pont judged it absent until 1986: about 11 years after its own pledge, and 12 after the mechanism was published.
• The threshold shaped timing. Precautionary action happened only under the lower threshold, where public pressure was strong (the US). International action waited for observed damage (pp. 79–80, 82).
• By 2001 the chapter itself uses a criminal-law standard (“beyond reasonable doubt”, p. 76). Notes’ own observation, not an argument Farman makes: reaching that standard took roughly 25 years from the 1974 mechanism, which illustrates what waiting for the highest standard of proof costs.

3. Industry incentives and behaviour#

• Concentration. 13 groups held about 75% of output, so a few actors could shape the transition (p. 81).
• From delay to managed transition. The trajectory ran: ‘wait and see’ (1975); some controls acceptable (1986); time for orderly reorganisation (1987, Farman’s reading of what the Protocol’s timing gave industry, not a stated industry position); profit from HCFC and HFC investment (1990) (pp. 80–81). In Farman’s account, the form of regulation (guidelines, open-ended timelines for substitutes) followed industry’s investment horizon (p. 81).
• Funding without acceptance. Substantial CMA research funding ran alongside Du Pont’s denial that reputable evidence existed (p. 80). No distortion of findings is alleged.
• Dispersive-by-design use. Aerosols sold a product whose whole content ends up in the environment (p. 82).
• Adaptive capacity. Once the direction was set, industry adjusted faster than expected (p. 81). Notes’ inference, not Farman’s claim: projected transition costs may overstate real ones. The chapter gives no cost figures.

4. Regulators and negotiators#

• Consensus as the operating norm. The voting rule was never used, the stricter-measures clause was formally invoked only once, the next meeting was always assumed to be easier, and terms were set by what was “thought practicable” (pp. 78, 81).
• The ratchet as a strength. Two-yearly reviews and immediate review after entry into force (p. 81) allowed tightening from no guaranteed recovery to recovery by 2050–60 (p. 78). Farman records these features but emphasises the slowness.
• Transatlantic divergence. The US favoured controls on uses, Europe a capacity cap. European NGOs were focused elsewhere and EC measures were token (p. 80).
• First-mover penalty and leakage. The US claimed its share fell from 46% to 28% because of its unilateral action, and it then refused to act alone again (p. 80). Maxwell and Briscoe (1997) independently estimate that the aerosol ban cost US producers about 20% of the global market (ext.).
• Managing credibility. Farman reads the timing of the signing as pre-emptive (p. 80). This is contested.
• Finance followed the technology path. The Fund financed the CFC-to-HCFC shift, locking in a second round of costs (p. 81). Developing countries received delayed schedules (p. 77).
• Maturing-regime problems. Illegal trade, and continuing CFC and halon consumption in Russia and other economies in transition (p. 81). Farman does not use the word “non-compliance”.

5. Publics, NGOs and media#

• Primed attention. The SST debate primed the US campaign that produced the 1977 action. European attention was on acid rain (pp. 79–80). Public attention is finite, and earlier controversies prepare the ground for later ones.
• NGO access and campaigning. NGOs were admitted to the UNEP committee (p. 79) and campaigned before London (p. 81).
• Framing. The chapter records the “expressive” coinage ‘ozone hole’ (p. 80) but does not analyse its effects.

6. Substitution, lock-in and the direction of innovation#

• Default substitution path. Rapid chemical-for-chemical substitution with industry-preferred HCFCs and HFCs became the default. Those substitutes were lightly regulated and given open-ended timelines (p. 81, fns 7–8).
• Forgone options. Avoidable-release reduction, halocarbon-free technology and energy efficiency, framed as serving ozone, climate and development together (p. 81). The editors echo this (p. 177).
• Lock-in. Through finance (Fund money went “largely” on replacing CFCs with HCFCs, p. 81), equipment and foam banks (p. 77), and producers’ investment in HCFC and HFC production (p. 81).

7. Time lags and irreversibility#

• Residence times. From about 5 to about 100 years (p. 77). Banks in equipment and foams keep releasing after production stops (791 kt of CFC-12 in 1995). For the unnamed halon, current releases come “predominantly” from developed-country banks (p. 77).
• Compounding delays. Loading, UV and latency compound, so harm peaks about 60 years after production does (Fig. 7.1).
• Practical irreversibility. Recognised too late, the problem could be neither avoided nor quickly relieved (pp. 82–83). The system is physically recoverable, but not on policy timescales.

8. Distribution#

• Benefits (refrigeration, insulation, solvents, aerosols) went to users and producers, unquantified (pp. 79, 82).
• Harms fall on future generations (latency) and on distant regions: released mainly in the north, with the worst loss over Antarctica (pp. 76, 79, 82). The chapter does not analyse the geographic asymmetry.
• Transition costs were partly socialised via the Fund, while producers could profit from substitutes (p. 81). That developed-country Parties pay into the Fund is external knowledge; the chapter says only that “the Parties” agreed to raise the money.

9. How key actors thought#

• Inventors and industry (1929): safety meant non-toxic and non-flammable in use. The stated selection criteria do not include environmental fate (p. 79); that it was not considered at all is an inference, supported by p. 82 (“not known what happens to CFCs” once released).
• Industry (1975–86): harm must meet the producer’s own evidential bar; research can be funded without conceding risk (p. 80).
• Conventional risk assessors (counterfactual 1965): inertness read as reassurance; decades without harm read as proof of safety; uncertainty became a research list; relative abundance used to dismiss mechanistic hints (p. 82).
• Scientists (early 1980s): theory dictated where to look, and agreement between models and observations was reassuring (pp. 80, 82).
• Meteorologists (1900): a “comfortable feeling” that the atmosphere was simple, which Farman calls “complacency” (p. 78). Reading it as a deliberate parallel for later expert overconfidence is the notes’ interpretation; Farman does not draw the link explicitly.
• Negotiators: consensus, practicability, “next meeting”, ready acceptance of industry’s transition (pp. 78, 81).
• Policy-makers: slow to register the move from ignorance to rudimentary understanding (p. 83).

10. Framing and language#

• ‘wonder chemicals’ (p. 82), used ironically.
• ‘wait and see’ against ‘reasonable expectation’ (p. 80).
• ‘reputable evidence’ (p. 80), the producer’s evidential gate.
• “so-called” essential uses (p. 77).
• ‘token gesture’ (p. 80).
• ‘psychological breakthrough’ (p. 81).
• ‘business as usual’, recast as an experiment (p. 82).
• ‘CFC episode’ (pp. 77–78), placing the whole affair in industrial history.

Farman also records others’ framings: a major success (p. 77); success, compromise, muddle, failure (p. 81). He positions himself against the triumphalist reading.

Transferable insights (technology-neutral)#

Ratings: strong means well evidenced here and corroborated; moderate means supported but by a single contrast or thin evidence; suggestive means illustrated but inferential or counterfactual; asserted means principle or opinion.

1. Safety where the product is used can be what creates hazard at system scale. Here inertness and low toxicity led to persistence and transport to a vulnerable part of the system. pp. 79, 82–83. Strong. The mechanism is fully documented and later confirmed; the broader generalisation is the author’s, but well grounded.
2. An early absence of observed harm is weak reassurance when pathways are delayed, remote or cumulative. Thirty years of “no apparent harm” (p. 82) and an early-1980s “no significant trend” (p. 80) both preceded serious damage. Strong.
3. Dominant models steer attention. Anomalies outside expectations get filtered by screening rules or doubted, and the decisive signal came from long-term observation not designed to test the model. pp. 80, 82. Strong (corroborated by NASA-side accounts; the data were flagged, not discarded).
4. Reasoning from relative quantity ignores amplification. Trace agents acting through chain or catalytic processes can matter disproportionately. pp. 79, 82. Suggestive. The 10,000-to-1 dismissal is a counterfactual, though the mechanism is established.
5. Whoever sets the evidential threshold largely sets the timing of action. A producer-defined ‘reputable evidence’ test postponed acceptance for about a decade; a statutory ‘reasonable expectation’ standard allowed action before proof. p. 80. Moderate. A clear contrast, but the ban’s legal basis is blurred (see Limitations).
6. A plausible mechanism brings limited action; observed, severe damage plus strong evidence of cause brings serious negotiation, though not necessarily decisive action. 13 years from mechanism to controls, against about 2 years from the observed hole to controls (pp. 79–82). Even then, Farman says the early negotiations failed “to act more decisively” and set terms by what was thought practicable (pp. 78, 81). Strong for the timing pattern. The hole’s causal role in 1987 is contested (Benedick).
7. Partial, use-specific or non-binding controls are easily offset. Growth elsewhere cancels restrictions on one use, and caps above actual output bind nothing. p. 80. Moderate. The chapter gives no numbers, but Maxwell and Briscoe (1997) independently report that by the mid-1980s growth in refrigeration, air conditioning and foams had more than offset the US aerosol-ban decline.
8. Unilateral first movers can lose market share, which then argues against further unilateral action. p. 80. Moderate. In the chapter this is a US claim that Farman reports, not his own finding. It is corroborated externally: Maxwell and Briscoe (1997) estimate the aerosol ban cost US producers about 20% of the global market.
9. Consensus-based incremental rule-making defers stringency, but built-in periodic review can turn a weak start into progressive tightening. p. 78, fn 6; p. 81. Moderate. The counterfactual of voting is untested, and the ratchet’s success partly cuts against Farman’s emphasis.
10. Phase-outs negotiated with concentrated incumbents tend to follow their preferred like-for-like substitutes. Transitional options are lightly regulated and open-ended, which seeds a second-round problem and repeated public costs. p. 81. Moderate in the text; strong in hindsight (HCFC acceleration in 2007; Kigali HFC phase-down in 2016). Whether a halocarbon-free route would have been as fast is unproven.
11. Framing a phase-out as “what replaces it one-for-one?” hides the option of reducing need and preventing avoidable releases. p. 81. Suggestive (the cited surveys are unnamed).
12. Ending production does not end the harm. Long residence times and stocks in use need managing in their own right. pp. 76–78; Figs 7.1–7.2. Strong.
13. Delays compound along a causal chain, so harm peaks decades after activity peaks and “success” is declared long before outcomes can be seen. Fig. 7.1, p. 76. Strong in structure; the figure is schematic.
14. Early precautionary action can depend on public mobilisation, which competes for finite attention and is primed by earlier controversies. pp. 79–80. Moderate. The evidence is a single contrast: an effective US campaign produced action, while Europe, with little public pressure, produced a largely token measure. Other drivers (for example consumer-market shifts, see Omissions) are not weighed.
15. A vivid public label for an abstract phenomenon can shift political momentum. p. 80. Suggestive here (the coinage is recorded, not its effects); external scholarship supports it.
16. Funding research is not the same as accepting its findings. Industry funded substantial research while denying reputable evidence existed. p. 80. Moderate (co-existence documented; influence not analysed).
17. Long-term systematic monitoring can be how surprises are found, yet it is hard to fund, and cannot be funded open-endedly at the level needed to uncover a major issue. p. 82. Moderate (one decisive case; the editors, pp. 172–173, read the same discovery as “essentially serendipitous”).
18. Stretched institutions, weak collaboration across groups and heavy assessment workloads delay the synthesis of scattered anomalies. p. 82. Suggestive.
19. Technologies are often deployed before the science needed to assess them exists. The key governance skill is recognising early when ignorance has become rudimentary understanding; acting on that recognition is implied rather than stated. pp. 79, 82–83. Moderate (strongly illustrated; the generalisation is asserted, and Farman frames it conditionally: “If there is to be a solution”).
20. No actor has a mandate for global, practically irreversible experiments, and continuing established practice can be such an experiment. p. 82. Farman names governments and multinational companies; “no actor” is the notes’ generalisation. Asserted (normative). The framing was not new in 2001: the US EPA administrator said in 1986 that society could no longer afford to “experiment” with the atmosphere, according to Maxwell and Briscoe (1997).
21. Agreements may be timed to pre-empt expected evidence, preserving negotiators’ credibility and giving industry transition time. p. 80. Suggestive. Inference from timing; contested; no documents.

Limitations, contestation and bias check#

Standpoint and genre#

• A participant’s reflective essay, not a documented history. The editors say Farman “discovered” the hole (p. 12). More precisely, he was lead author of the three-author 1985 paper (Farman, Gardiner and Shanklin; p. 83). The chapter is short, compressed and lightly referenced.
• Candour reduces hindsight bias. The surprise to his own team (p. 82) and the modest 1965 counterfactual (p. 82) make the chapter less hindsight-driven than much precautionary writing. Farman labels his retrospective judgements as such.
• An unaddressed blind spot. He does not discuss his own team’s delay before publishing the Antarctic signal. Benedick (2005 testimony) says the British group “initially hesitated to publish” because the findings seemed too fantastic. Other accounts describe several years of low readings and instrument checks; I have not verified those details. This bears directly on his lesson about recognising anomalies. Farman’s candid “including the authors” (p. 82) partly acknowledges it.

Contested interpretations#

• Was Montreal precautionary?
• Farman: no; events overtook the negotiators, with pre-emptive timing (p. 80).
• Benedick (US Senate testimony, 28 September 2005, a primary statement of his view; checked against the PDF): the Protocol imposed substantial short-term costs to guard against speculative future dangers that rested on scientific theories rather than proven facts; he calls this “preventive action on a global scale”. The Antarctic hole was widely seen during the negotiations as an anomaly that did not fit the models. The models then predicted no significant global depletion for at least 20 years. The decisive evidence (the airborne expedition results and the March 1988 Ozone Trends Panel) came after signing.
• Assessment: partly reconcilable. The hole created political urgency without being part of the formal scientific case. Maxwell and Briscoe (1997) say the hole “was a political trigger even if it was not a scientific one”. Farman’s reading of motives rests on inference from timing.
• The HCFC/HFC pathway.
• Farman does not weigh the trade-off. The transitional substances may have enabled the fast industry response he himself acknowledges (p. 81) and the quick exit of CFCs from developed countries.
• His halocarbon-free counterfactual was later partly vindicated, but is untested for speed, safety (for example flammability) and feasibility sector by sector.
• The supporting “technical surveys” are uncited (p. 81).
• The US SST decision (p. 79). Congress ended SST funding in 1971 over several concerns, including cost, sonic booms and noise; ozone was only part of the debate. Farman’s “heeded” framing overstates the role of ozone concerns.

Factual and editorial slips#

None of these changes the argument, but they matter before citing. - HCFC-22 is called “a hydrofluorocarbon” (p. 76). - The halon passage (p. 77) is missing the halon’s name; it is probably halon-1301. - Footnote 6 (p. 78) quotes the 1987 voting rule (two-thirds representing at least 50% of consumption). The London Amendment (1990) had replaced this with a two-thirds majority that must include majorities of both Article 5 and non-Article 5 Parties (current wording confirmed on the Ozone Secretariat site in the audit; attributing the change to London is from knowledge of the amendment text, not that page). The point that no vote had been taken still stands. - “Invoked only once, by the EU in March 1991” (fn 6). This probably refers to EC Regulation 594/91. But national measures stricter than the Protocol existed at the same time (for example, Germany’s CFC-Halon Prohibition Ordinance of May 1991), so the claim holds only for formal invocation. Treat it as uncertain. “European Union” is also an anachronism: in March 1991 the actor was the European Community (the EU dates from November 1993). - The US aerosol ban (p. 80). It was finalised in March 1978 under the Toxic Substances Control Act and the Food, Drug and Cosmetic Act, effective 1979 (secondary sources; verify in 43 FR, March 1978). The 1977 Clean Air Act amendments separately set the precautionary standard, in the statutory wording “may reasonably be anticipated” (section 157, as quoted by Benedick 2005). The phrase Farman quotes comes via the 1987 EPA notice and appears to reflect legislative history; I could not verify it verbatim. The standard is real, but “explicitly states” overstates it, and the link between the Act and the ban is blurred. Table 7.1 repeats the link (p. 83). - Minor: “Midgely” (Midgley); “Eric” (Erik) Rotheim; Teisserenc de Bort’s announcement is usually dated 1902, not 1901; “Antartica” (Table 7.1). - Slaper et al. (1996) is cited in the Fig. 7.1 caption but not listed in the references. - Table 7.1 versus text: the table says refrigerators (p. 83); the text says foam blowing (p. 80). - Table 7.1’s 1987 entry describes the original Protocol as phasing out ozone-depleting substances; it set freezes and reductions (see the Table 7.1 notes above). Table 7.1 and Fig. 7.1 are EEA-credited, so these may be editorial rather than authorial slips. - Minor (p. 81): “Scientific and Economic Assessment Panels”. The Protocol’s bodies are the Scientific Assessment Panel and the Technology and Economic Assessment Panel (ext.).

Omissions that matter for using the chapter as a lens#

• Non-linearity. No account of polar stratospheric cloud chemistry or threshold behaviour. Benedick (2005) stresses that chlorine tripled with no effect until a threshold was crossed. Fig. 7.2’s 2.5 ppbv line hints at this, but the complex-systems lesson is not drawn.
• 1988 and industry’s research choices. The Ozone Trends Panel and Du Pont’s March 1988 pledge to phase out CFCs entirely are absent. So is Du Pont’s early-1980s shift of R&D away from commercialising substitutes. That shift came as regulatory pressure eased, though Du Pont attributed it to market conditions (the substitutes were too costly) rather than corporate policy; the company moved decisively back towards substitutes in 1986 (Maxwell and Briscoe 1997). That history would add context to Farman’s account of industry’s response (pp. 80–81).
• Consumer and market pressure. A major US consumer-products firm dropped CFC propellants voluntarily in 1975, and falling consumer demand preceded the ban (Maxwell and Briscoe 1997; rechecked in the second audit, which also confirmed their “political trigger” wording, the 20% market-share estimate, the mid-1980s offset and the patent-expiry denial). Maxwell and Briscoe put the CMA-coordinated research contribution at “some $3–5 million” over several years, a figure to set beside Farman’s “substantial financial support” (p. 80).
• European producers. Farman criticises European governments’ measures (p. 80) but not European producers. Benedick (2005) names the EC, influenced by its major producers, as the main obstacle. An EEA-published chapter naming Du Pont but no European firm may reflect venue rather than bias, but readers should know the other account exists.
• Numbers. No figures for the benefits of CFCs, the costs of control or the health gains; Fig. 7.1 is schematic.
• Geography of harm. Mostly northern emissions against southern and polar harm is noted only in passing (pp. 79, 82).

Precautionary framing and editorial overreach, fairly assessed#

• Against the book’s own framing. The report features Montreal’s preamble on “precautionary measures” (Table 1.2, p. 14) and lists the 1977 US aerosol ban as an example of ‘precautionary prevention’ (Table 1.1, p. 12). Farman agrees about 1977 but is less flattering about the rest: unequivocal precaution was used once, weakly, and then largely abandoned until severe depletion and strong evidence arrived (pp. 80, 82).
• The synthesis goes further than the chapter:
• Satellite data. p. 173 says the results were considered suspect and ‘set aside’. Farman (p. 82) says they were flagged and being re-examined. Later NASA-side accounts support Farman (Schmidt 2017): the data were quality-flagged, not deleted; NASA’s processing team began investigating the low October 1983 values in August 1984 and had a conference abstract in by December 1984; maps were prepared but presented publicly only after Farman’s paper. Benedick (2005) repeats the “automatically reject” version.
• How the discovery happened. pp. 172–173 call the discovery “essentially serendipitous” and “essentially by accident”, a by-product of research done for other purposes. Farman (p. 82) stresses systematic long-term measurements begun for scientific exploration. The two framings draw different lessons: luck versus sustained observation.
• Regulatory response. p. 173 says the 1974 mechanism “did not prevent regulatory neglect” until firm empirical evidence arrived. Farman agrees that a plausible case brought only limited action (p. 82). But “neglect” erases the 1977 US/Nordic bans, the 1977 UNEP plan and the 1980 EC decision, which he records as the one unequivocally precautionary episode (pp. 79–80).
• Foreseeability. p. 170 says novelty, persistence and ubiquity were knowable “at the outset” as warnings. Farman is more cautious: the persistence of inert chemicals took a long time to be realised (p. 83), and a 1965 assessment would have found nothing (p. 82).
• Not an overstatement (corrected in audit). p. 187’s “near monopoly” refers to reliance on a single technology to meet a need, as against “several competing technologies”; it is not a claim about corporate market share. It does not conflict with the chapter’s 13 groups holding about 75% of production (p. 81). The editors’ diversity argument (p. 187) and their criticism of “second-generation CFC substitutes” (p. 177) are consistent with Farman’s substitution critique.

What later evidence says (brief; full hindsight checking belongs to a later stage)#

• Recovery. The WMO/UNEP 2022 assessment projects that total column ozone returns to 1980 values around 2066 over Antarctica, 2045 over the Arctic and 2040 near-globally (60°N–60°S). That is broadly consistent with the chapter’s 2050–60 return of tropospheric equivalent chlorine to 2.5 ppbv (p. 78), though the metrics differ. The Antarctic lags.
• Halon-1301. Its rate of change was “indistinguishable from zero” over 2016–2020 (WMO 2022), consistent with Farman’s projected rise until about 2020 (p. 77). To my knowledge, destruction of banks was never made mandatory (to verify).
• Substitutes. The HCFC phase-out was accelerated in 2007, and the Kigali Amendment (2016, in force 2019) phases down HFCs, avoiding an estimated 0.3–0.5 °C of warming by 2100 (WMO 2022). This bears out the concerns behind the p. 81 critique and footnote 7. It does not show that Farman’s halocarbon-free alternative was feasible in 1990.
• Climate co-benefit. Velders et al. (2007) found that the Montreal Protocol’s climate benefit already exceeded Kyoto’s first-period target. WMO 2022 estimates it avoids about 0.5–1 °C of warming by mid-century, relative to an extreme scenario of uncontrolled ODS growth. This supports Farman’s linking of ozone, climate and technology choice (pp. 81–82).
• Enforcement. Unexpected CFC-11 emissions after 2012 came largely from eastern China and fell after 2018. Carbon tetrachloride is declining more slowly than expected, possibly because feedstock-related emissions are underestimated (WMO 2022). Both bear out the illegal-trade concern (p. 81) and the flag on feedstock exemptions (p. 77).
• Health. A US EPA model (2020) estimates that full implementation prevents about 443 million skin cancers, 2.3 million skin-cancer deaths and 63 million cataracts among people in the US born 1890–2100. These are model estimates, but they support Fig. 7.1’s logic.
• The Multilateral Fund grew far beyond the USD 1 billion on p. 81, to roughly USD 4–5 billion in cumulative contributions by the 2020s (search-derived; verify).
• Du Pont. Maxwell and Briscoe (1997) argue that its support for international regulation was also commercial: CFC margins had been poor since the late 1970s and substitutes promised specialty-chemical margins. This corroborates p. 81. Du Pont officials denied that patent expiry was a factor.

External sources consulted: - WMO/UNEP, Scientific Assessment of Ozone Depletion: 2022, Executive Summary: https://csl.noaa.gov/assessments/ozone/2022/executivesummary/ - R. E. Benedick, Senate EPW testimony, 28 September 2005: https://www.epw.senate.gov/public/_cache/files/1/0/1094f8dd-dae7-45cd-b5a1-b9c018272318/BF9D594B66EBA773D15F23EC2FEC547786CB6ADB4C2DD1862C0C90B6D44D8B5A.092805benedick-testimony.pdf - G. Schmidt, “What did NASA know? and when did they know it?”, RealClimate, 2017: https://www.realclimate.org/index.php/archives/2017/12/what-did-nasa-know-and-when-did-they-know-it/ - J. Maxwell and F. Briscoe, Business Strategy and the Environment 6 (1997) 276–286: https://gwern.net/doc/politics/1997-maxwell.pdf - Montreal Protocol Art. 2 (current text): https://ozone.unep.org/treaties/montreal-protocol/articles/article-2-control-measures - G. J. M. Velders et al., PNAS 104 (2007) 4814–4819: https://www.pnas.org/doi/10.1073/pnas.0610328104 - US EPA AHEF report (2020): https://www.epa.gov/sites/default/files/2020-04/documents/2020_ahef_report.pdf; Ozone Secretariat summary: https://ozone.unep.org/montreal-protocol-likely-avert-443-million-skin-cancer-cases-united-states - Germany’s FCKW-Halon-Verbots-Verordnung (1991): https://www.osti.gov/etdeweb/biblio/5666280

Notable passages#

Close paraphrases with pages; one verbatim fragment. Exact wording is retrievable from the PDF.

1. p. 76. Causation is established “beyond reasonable doubt” by the WMO assessments.
2. p. 77. Ending releases is not the end of the story: CFC-12 will be at least 37% of today’s level in 2100, and banks keep leaking.
3. p. 78. Early failure to act decisively “can be seen perhaps” as ‘consensus before effectiveness’, with the next meeting always expected to be easier.
4. p. 80. The 1977–80 measures “appear to be” the ‘first, and last’ unequivocal use of precaution, and the only action before the evidence was compelling.
5. p. 80. Du Pont’s ‘reputable evidence’ pledge (1975) and its denial until 1986, set against ‘wait and see’ and the ‘reasonable expectation’ legal standard.
6. p. 80. Montreal’s timing makes sense only as a ‘pre-emptive move’ for the negotiators’ credibility and industry’s orderly reorganisation.
7. p. 81. “In my view” the open-ended HCFC/HFC approach was ‘deeply flawed’; radical, halocarbon-free change should have been stimulated from the start.
8. p. 82. A 1965 risk assessment would have found no known grounds for concern, and early hints would “surely” have been dismissed on a 10,000-to-1 abundance ratio.
9. p. 82. Neither governments nor multinationals have a mandate for global experiments, even ‘business as usual’.
10. p. 83. Short-term safety “appears to demand” non-reactivity, which means persistence. Policy-makers must recognise “when ignorance has been replaced by understanding, however rudimentary” (verbatim).

Open questions#

1. What drove the September 1987 timing? Farman’s pre-emptive-move reading against Benedick’s precaution-under-uncertainty reading. Primary records could help: Benedick’s Ozone Diplomacy, UNEP negotiating documents, and the Cagin and Dray pages Farman cites.
2. How strong was the 1990 case for going halocarbon-free? What were the “technical surveys” (p. 81)? Which sectors could have skipped HCFCs and HFCs, at what cost and safety risk? Did the transitional substances speed up the CFC exit?
3. Did the ‘reasonable expectation’ standard actually drive the 1977–78 US action, or did consumer-market shifts and other statutes do most of the work?
4. Did CMA funding (p. 80) shape the research agenda, the speed of consensus or the framing of uncertainty? Benedick (2005, checked against the testimony text) claims that in 1985 the UK government stopped financing the British Antarctic mission that had found the hole, for “political motives”, and that the US CMA filled the gap. Only Benedick makes this claim; it needs checking against British Antarctic Survey or UK government records.
5. Why are the non-linear, threshold dynamics of polar ozone loss absent from the lessons? They are arguably the most generalisable complex-systems lesson of the case.
6. How should the p. 83 recommendation be put into practice? Which signals (a mechanism, measured persistence, measured ubiquity) should mark the point where understanding has arrived? The chapter implies 1973–74; the editors (p. 170) imply earlier.
7. Were the chapter’s projections borne out? A halon peak around 2020, equivalent chlorine back to 2.5 ppbv by 2050–60, developing-country cessation by 2010. WMO 2022 suggests broadly yes, with caveats (CFC-11, carbon tetrachloride). This is for the hindsight stage.
8. Is the European role treated gently in this EEA chapter compared with other accounts (Benedick 2005)?
9. Does the regime’s eventual strength argue for consensus with periodic review, against Farman’s emphasis on slowness (p. 78)?

Audit log#

Independent audit against the text extract, the rendered PDF (pp. 12, 14, 76–83, 169–173, 177, 187, 195), Benedick’s 2005 testimony, Maxwell and Briscoe (1997), Schmidt (2017), the Ozone Secretariat Article 2 page and the WMO 2022 Executive Summary.

• Stance: restored Farman’s hedge (“appear to be” the only unequivocal precaution), added the “psychological breakthrough” credit and the Beijing “glimpse”, and marked his consensus critique as hedged.
• Benedick: changed “reports and rejects” to “offers ‘a simpler interpretation’ against”; noted the chapter has no panels or boxes.
• Added an attribution caveat: Fig. 7.1 and Table 7.1 are credited “Source: EEA”, so their wording (including the ‘success’ scare quotes) is not necessarily Farman’s.
• Sources: added Goody 1954 and the IISD web report (fn 9); noted that the Du Pont NYT statement is cited in the text only.
• Fig. 7.1: corrected the reading of the caption, which credits stopping CFC production with averting much more skin cancer; the scare quotes qualify rather than dismiss.
• Fig. 7.2: corrected the crossing of the 2.5 ppbv line from “around 1980” to the late 1970s (about 1977; chlorine alone about 1981), checked visually.
• p. 78 key claim: flagged that the projection sequence (no guaranteed return until 1995) is unsourced in the chapter.
• p. 78 diagnosis: restored “can be seen perhaps” and “in retrospect it seems”.
• 7.4: added Farman’s “first significant move” label; restored “appear to be”; noted that “these measures” includes the token EC decision.
• 7.4 and throughout: the 46% to 28% US share fall is reported by Farman as a US claim, now labelled so (notes, mechanisms, insight 8).
• 7.5: added “astutely designed” to the pre-emptive-move reading.
• 7.6: restored “might reasonably have been asked” in place of “should have been questioned”; added that recession and war held back early growth.
• 7.6: corrected footnote 10 (nothing suggests influence on workers “in the 1960s and early 1970s”, not “no one before the 1970s”); added the conditional (“if it had been recognised” that chlorine could be released); “likely” corrected to “surely”; “shows” corrected to “suggests”.
• 7.6: added a reading note on how the chapter implicitly answers its own “should/could it have been avoided” questions.
• Closing recommendation: added the “If there is to be a solution” conditional; marked “and act then” as implied.
• Table 7.1: flagged that the 1987 entry’s “phasing out” overstates the original Protocol, which set freezes and reductions.
• Timeline: removed “dramatic observation produced decisive action”, which contradicts p. 78 (“failure to act more decisively”); replaced “science barely existed” with a p. 79-consistent statement; widened “1987–89” to 1987–89/90.
• Lessons 1, 3, 6: restored hedges (“appear to be”, “suggests”, “perhaps”, “surely”, “just possibly”).
• Recommendations: marked “leapfrog” as the notes’ gloss.
• Mechanisms: fixed the arithmetic (1975–86 is about 11 years after the pledge, 12 after the mechanism); relabelled the “beyond reasonable doubt took 25 years” point and the transition-cost point as the notes’ own inferences, not Farman’s.
• Mechanisms: replaced “non-compliance” (not in the source) with “continuing consumption”; replaced “Fund paid for HCFC capacity” with the source’s “replacement of CFCs by HCFCs”; marked the developed-country funding of the Fund as external.
• Actor thinking: marked “environmental fate not considered” and the meteorologists-as-parallel reading as inferences; added Farman’s word “complacency”.
• Insight 6: reworded to “serious negotiation, though not necessarily decisive action”.
• Insight 7: added Maxwell and Briscoe’s corroboration of the foam/refrigeration offset.
• Insight 8: labelled as a US claim; verified Maxwell and Briscoe’s 20% figure.
• Insight 17: “is how” softened to “can be how”, with the editors’ “serendipitous” counter-reading added.
• Insight 20: added the 1986 EPA “experiment” framing (Maxwell and Briscoe) for context.
• Limitations: added Benedick’s statement that the British group “initially hesitated to publish”; checked the Benedick paraphrase against the testimony PDF; quoted Maxwell and Briscoe’s “political trigger” wording exactly.
• Omissions: corrected the “halt to alternatives research” account (Du Pont attributed the early-1980s R&D shift to market conditions; it resumed in 1986).
• Framing: “precautionary success” corrected to Table 1.1’s “precautionary prevention”; “overwhelming” evidence corrected to “strong”.
• Main correction: removed the claim that p. 187’s “near monopoly” overstates market concentration. The editors mean reliance on a single technology, not firm market share, so the passage is consistent with the chapter.
• Editors versus Farman: added p. 173’s “regulatory neglect” (which erases the 1977–80 measures) and the “serendipitous/by accident” versus “systematic measurements” contrast (pp. 172–173 v. p. 82).
• Satellite data: made the NASA account more precise from Schmidt 2017 (data flagged, not deleted; investigation from August 1984; maps made public only after Farman).
• Later evidence: specified that WMO 2022 recovery dates are for total column ozone; halon-1301 “indistinguishable from zero”; carbon tetrachloride wording corrected; Montreal’s avoided warming is relative to an extreme scenario; “substantially vindicates” softened to “bears out the concerns”, noting that it does not prove the halocarbon-free route was feasible.
• Slips: added the Table 7.1 1987 entry and the misnamed assessment panels (p. 81); clarified what the Ozone Secretariat page does and does not confirm about fn 6.
• Open question 4: stated Benedick’s CMA/UK-funding claim precisely.
• Quotation policy: updated to cover the short verbatim hedges added.
• Checked the notes for any mention of technologies or companies outside the report: none found. External firms are anonymised, as before.

Second audit pass (2026-09-25)#

Independent re-check against the full text extract; the PDF pages rendered as images (pp. 76–78, 80–83: Figs 7.1–7.2, Table 7.1 and layout, confirming no panels or boxes); the PDF text layer for the editors’ pages (pp. 2, 12, 14, 169–173, 177, 187, 195); and the text of Maxwell and Briscoe (1997). Earlier corrections were confirmed; the changes below are additions or refinements.

• Verified: Fig. 7.1 readings (not to scale; production peak about 1989; extra skin cancer peaking about 2050), Fig. 7.2 crossings and peaks, all 13 Table 7.1 entries, the author note, and the editors’ wording on pp. 12, 14, 169, 170, 172–173, 177 and 187.
• p. 77 halon gap: confirmed in the PDF text layer as well as the print; added the reasoning for the halon-1301 inference (lifetime).
• 7.1: “answers through three sketches” corrected to “may help to put these questions in perspective” (Farman does not answer directly).
• Footnote 6: added Farman’s own inference that negotiators were “quite content” with a consensus pace; flagged “European Union” (March 1991) as an anachronism for the European Community (notes and digest).
• 7.2: “left no residue” corrected to “no objectionable residue”; added “most powerful oxidant known” and “now largely ignored in textbooks”.
• 7.3: “patented in 1926” corrected to “applied for a patent”; “reasonably well characterised” replaced by the source’s “quite comprehensively investigated”, considering the resources available; added Fabry and Buisson and the two profile methods.
• 7.4: EC Decision 80/372 “froze… required” corrected to “aimed to freeze… and to cut”; heading changed from “achieved little” to Farman’s “not easily seen”; added “little public pressure” in Europe.
• 7.4: Du Pont’s 1975 pledge restated more precisely (a health hazard through ozone depletion; stop the “offending compounds”); added Farman’s description of Du Pont as the first and largest producer.
• 7.4: “The US would not act alone again” corrected to “by this time (1985) was not prepared to act unilaterally”.
• 7.5 Beijing: panels asked about the “development and marketing” of new ozone-depleting chemicals, not only marketing.
• 7.6: 1965 counterfactual corrected to “no known grounds for concern”; added the omitted point that the assessment might have noted CFCs’ unknown atmospheric fate but discounted it against 30 years without apparent harm (notes, lessons, notable passages, digest).
• 7.6: added a check that “serious negotiations did not begin until” severe depletion leans on “serious”, since protocol talks were authorised in March 1985, before the Antarctic paper (p. 80) (notes and digest).
• 7.6 and 7.6 close: restored the hedges “appear to be reasonable” and “appears to demand”; added “stark warning” and “deep-seated paradox” (notes, lessons, notable passages, digest).
• Timeline: “chlorine-catalysed” corrected to the source’s “photosensitised by chlorine”; the Sep 1987 row now says NOZE II and AAOE press releases were “expected”, not that signing preceded “results”; the 1930 knowledge entry now cites p. 82.
• Recommendations and insight 20: “No actor has a mandate” restored to Farman’s “neither governments nor multinational companies”; “no actor” marked as the notes’ generalisation (notes and digest).
• Mechanisms: “stricter measures rarely invoked” corrected to “invoked only once”; “banks dominate the atmospheric tail” softened to what p. 77 says; 1987 “orderly reorganisation” marked as Farman’s reading of the Protocol’s timing, not a stated industry position; “consolidating versus noticing” marked as the notes’ gloss.
• Insight 14: “depends on” softened to “can depend on”, with a note that it rests on a single US/Europe contrast and ignores market drivers.
• Insight 19: acting on recognition marked as implied; added the conditional “If there is to be a solution”.
• Limitations: “Farman discovered the hole” attributed to the editors (p. 12), noting that he was lead author of a three-author paper (also in the digest).
• Omissions: rechecked Maxwell and Briscoe against their text (the 1975 consumer firm, “political trigger”, the 20% share, the offset, patent expiry); added their $3–5 million figure for CMA-coordinated research funding.
• Digest: 1929–34 sketch reworded to avoid implying ozone science was absent; the Vienna Convention’s protocol mandate added; hindsight line tied to specific concerns (fn 7; Beijing illegal trade) rather than “his warnings”.
• Quotation policy: updated for the added short hedges; all new verbatim fragments are five words or fewer.
• Re-scanned the notes and digest for contemporary technologies or companies outside the report: none. “Software” and “Internet” appear only as the source uses them (p. 82).