LL1-07 hindsight check: 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, pp. 76–83. Author: Joe Farman.

Check window: publication (2001) to late September 2026. Checked: 25 September 2026.

Method note. The check was done in two passes on the same day. In the first pass general web search was unavailable (the session’s search budget was exhausted), so sources were retrieved directly from primary repositories:

• the WMO/UNEP Scientific Assessment of Ozone Depletion: 2022 (executive summary and Chapters 1 and 7, via the NOAA Chemical Sciences Laboratory mirror);
• its Twenty Questions and Answers (2022 update) and the 1994 assessment’s executive summary;
• NOAA’s Ozone-Depleting Gas Index;
• the Ozone Secretariat’s decision texts, treaty text and ratification table;
• the Multilateral Fund, World Bank project documents and the TEAP 2022 assessment;
• European Commission regulation pages, the UN Treaty Collection, US EPA and the White House;
• peer-reviewed papers via Europe PMC, Crossref and publisher pages.

In a second, resumed pass, web search was available again. It was used to close gaps the first pass had flagged:

• Benedick’s 2005 Senate testimony, the “2005 testimony” the digest refers to (Claim 8);
• the original 1987 text of the Protocol’s preamble (Claim 8);
• the World Bank’s 2007 press release on China’s CFC and halon plant closures, and China’s halon phase-out plan (Claim 6);
• DuPont’s March 1988 commitment and the industry coalition’s September 1986 shift (Claim 9);
• the 2024 and 2025 Meetings of the Parties, WMO’s September 2025 Ozone and UV Bulletin and the NASA/NOAA account of the 2025 ozone hole (Claims 1, 3 and 7).

The next quadrennial WMO/UNEP assessment (2026) has not been published. NOAA says the core report “will be released at the end of 2026” and the Executive Summary is expected “late 2026” (NOAA CSL, Scientific Assessment of Ozone Depletion: 2026 page, accessed 25 Sept 2026). The 2022 assessment is therefore still the latest full consensus statement. It is supplemented here by papers and official bulletins from 2023 to 2026.

The full 2013 Late lessons volume was retrieved from the EU Publications Office, so its Annexes 2 and 3 could be checked directly. Where a point rests on a source I could not open, I say so.

Overview#

What holds. The chapter’s physical forecasts, and its central warning about persistence, have aged well.

• CFC-12 persistence.
• The chapter’s forecast: CFC-12 will still be at least 37% of its 2001 level in 2100.
• The WMO 2022 baseline puts it at about 44% (236 ppt against 543 ppt), partly because banks turned out larger than assessed.
• Halon-1301.
• The chapter’s forecast: the long-lived halon would rise until about 2020 unless banked stocks were destroyed.
• It rose until about 2015, then stopped changing measurably over 2016–2020 at about 3.3 ppt.
• No destruction requirement was ever adopted.
• Methyl chloroform. It fell to about 1% of its peak by 2020, in line with the chapter’s “about 2% after 20 years”.
• The transition critique has been largely vindicated by the Parties’ own later actions.
• In 2007 they accelerated the HCFC phase-out. The same decision told the Multilateral Fund to change its rules on “second conversions” (paying again to convert firms it had already moved to HCFCs), which bears out the chapter’s point that the Fund would have to pay twice.
• In 2016 they brought HFCs under control through the Kigali Amendment, which had 174 parties by August 2026.
• Halocarbon-free refrigerant technology now dominates new domestic refrigeration worldwide.
• The illegal-trade warning came true in a harder form. Unreported CFC-11 production in eastern China after 2012 was detected by long-term atmospheric monitoring and had largely stopped by 2019.

What needs qualifying.

• The recovery date has slipped. The chapter expected the late-1970s chlorine level to return in 2050–2060. The latest assessment gives:
• mid-latitude EESC (equivalent effective stratospheric chlorine, the standard index of ozone-destroying chlorine and bromine) back to its 1980 level in about 2066 under its revised method;
• polar EESC in about 2087;
• Antarctic total ozone back to 1980 values around 2066.

NOAA’s index, which uses an older method, still projects around 2045 for mid-latitudes and around 2070 for Antarctica. The return date depends on the metric. Projected dates have also tended to recede as larger banks, feedstock losses and unreported production were found. Recovery itself has now been statistically detected (Solomon et al., 2016; Wang et al., 2025). The 2025 Antarctic hole was the fifth smallest since 1992, and NASA and NOAA still describe full recovery as expected “around the late 2060s”. - The negotiating history needs correcting. The chapter says the 1990 and 1992 amendments gave “no guarantee” of recovery and that only the 1995 adjustments secured a projected return. WMO’s 1994 assessment already projected recovery around 2045, and WMO’s 2022 retrospective dates the first projected long-term decline in EESC to the 1992 Copenhagen Amendment. - A dating slip. The chapter says halon production in China, India and Korea was “due to cease in 2002”. For developing countries 2002 was the treaty’s freeze year, and China ended halon production for emissive uses around 2006–2007, closing its last CFC and halon plants on 1 July 2007. - Skin cancer. Later modelling by the group behind the chapter’s source supports the shape of Fig. 7.1: an excess peaking around mid-century and then falling. But it puts the peak excess at no more than about 4% of local incidence. The burden remains a modelled quantity. It is not observed in cancer registries, where behaviour dominates the trends.

What is contested. Whether the 1987 Protocol was “precautionary” remains a dispute about definitions.

• For Farman’s reading. The Protocol’s own preamble describes the 1977–80 national and regional measures as “precautionary measures”, which is his framing.
• Against it. The same preamble (in its original 1987 wording) describes the Protocol’s own controls as precautionary. Benedick restated that reading in 2005 Senate testimony: the arguments for control “rested on unproven scientific theories”, and most scientists then saw the Antarctic hole as an anomaly. The Multilateral Fund now describes the 1985 Vienna Convention as the first major treaty to apply the precautionary principle.
• Where the two accounts agree. Both accounts put the decisive evidence after the signing: the Antarctic expedition’s preliminary results about two weeks later, and the Ozone Trends Panel in March 1988. They disagree about what moved the negotiators before then.

What the chapter did not foresee.

• Exemptions became leaks.
• Emissions from ODS feedstock use run at about 3.6% of production, against the 0.5% assumed when feedstocks were exempted. Left unchecked, this could delay mid-latitude recovery by about 7 years (Reimann et al., 2026).
• The Parties’ 2024 response (Decision XXXVI/5) asks for emissions to be minimised but leaves the exemption intact, and the 2025 meeting again reached no agreement on feedstocks.
• Several CFCs rose in 2010–2020 as by-products of HFC production.
• HFC-23 emissions rose while reported abatement said they should fall.
• The halon bank became a resource to conserve.
• Banked halons turned from a stock to be destroyed into a scarce supply that civil aviation still depends on.
• The Parties now worry about shortage.
• EU law now discourages destroying recovered halon.
• Uncontrolled ozone-depleting emissions. Nitrous oxide (N2O) and dichloromethane, which the Protocol does not control, became the main uncontrolled ozone-depleting emissions.
• Russia and the economies in transition. This problem, which the chapter lists as unresolved, was solved within about a year: Russia closed CFC production on 20 December 2000.
• The contrast with climate policy sharpened. The chapter measured Montreal’s lessons against climate policy.
• The ozone regime became universal (198 parties), and the US ratified Kigali in 2022.
• The US never ratified Kyoto, withdrew from Paris with effect from 27 January 2026, and on 7 January 2026 directed withdrawal from the UNFCCC (the UN climate convention) and the IPCC.

Annex 3 of the 2013 report (Late lessons II, p. 728) carries an “Ozone layer update”.

• What it is. A summary of WMO’s Twenty Questions: 2006 Update.
• What it says. Recovery is expected “near the middle of the 21st century”.
• What it does not do. It does not assess any of Farman’s claims.

Elsewhere in the 2013 volume:

• Annex 2 (p. 708) reprints the 2001 summary and timeline unchanged, including “restore levels of chlorine by 2050–2060”.
• Chapter 23, Box 23.3 (pp. 575–576) uses the ozone case to estimate the benefits of early action, citing a return to pre-1980 levels “between 2050 and 2075” (UNEP 2009).
• Chapter 26 contains an error. A footnote there (p. 631) misnames Farman (“Joe Forman”). It also says he returned to Antarctica three times and published under pressure from his funders. Neither detail is in his chapter and neither is sourced, so the footnote should not be relied on.

Claim 1: When chlorine loading returns to its late-1970s level, and which agreement secured that#

Original claim (p. 78; Fig. 7.2 p. 77; Table 7.1 p. 83).

• The metric. The chapter tracks “equivalent chlorine” in the troposphere: chlorine plus 58 × bromine. It treats 2.5 ppbv, first reached in the late 1970s, as the significant level.
• The forecast. Under the 1997 amendments, a return to that level is expected between 2050 and 2060.
• The history. Under the 1987 Protocol and the 1990 London and 1992 Copenhagen amendments there was “no guarantee” of any return. Only the 1995 Vienna adjustments secured a projected return, and the 1997 amendments brought it “a few years nearer”.
• The diagnosis. Negotiators put consensus before effectiveness.

Subsequent developments.

• The current projections are later, and depend on the metric.
• WMO 2022 projects total column ozone back to 1980 values “around 2066” over the Antarctic, around 2045 over the Arctic and around 2040 for the near-global average (60°N–60°S) (WMO/UNEP 2022 Executive Summary, pub. Jan 2023).
• For EESC, the 2022 baseline has mid-latitude EESC returning to 1980 levels “at the beginning of 2066” and polar EESC in 2087 (WMO/UNEP 2022, Ch. 7, §7.4.3.1).
• The same chapter notes that its revised EESC formulation (Engel et al., 2017) by itself moves the mid-latitude return more than a decade later than the older method used in the main tables of the 2018 assessment.
• Why the dates moved, beyond method. WMO 2022 gives these reasons (Ch. 7, Summary):
• the return is delayed by 4 years (mid-latitude) and 7 years (polar) relative to the 2018 baseline, “due mainly to a larger assessed CFC-11 bank” and a larger CFC-12 bank;
• unreported CFC-11 emissions over 2012–2019 add about one more year;
• eliminating all future long-lived ODS emissions, which are dominated by releases from banks, would bring the returns forward by 16 and 19 years.
• Feedstock emissions could add a further delay. A 2026 analysis finds that feedstock emissions run at “typically 3.6% of production”, not the 0.5% assumed when feedstocks were exempted. Left unchecked, they could delay mid-latitude recovery by 7 (6–11) years (Reimann et al., Nature Communications, 16 Apr 2026).
• NOAA’s older index gives earlier dates. NOAA’s Ozone-Depleting Gas Index uses the older fixed-lag approach. Its 2024 values show mid-latitude reactive halogen 55% of the way back to the 1980 benchmark and Antarctic 28% of the way back. It projects a return around 2045 for mid-latitudes and around 2070 for Antarctica (NOAA ODGI, 2024 report, fall 2024).

Farman’s 2050–2060 window therefore falls between the NOAA mid-latitude date and the WMO 2022 dates. His tropospheric 2.5 ppbv threshold roughly corresponds to the 1980 stratospheric benchmark, given a transport lag of about 3 years. - Recovery is now detected, not just projected. - Solomon et al. identified the first fingerprints of September “healing” of the Antarctic ozone layer (Science, 30 June 2016). - Wang et al. used formal pattern-based detection and attribution. They found “robust statistical and physical evidence” that ODS reductions under the Protocol are producing the beginning of Antarctic recovery (Nature, 5 Mar 2025). - WMO 2022 notes that 2017–2020 total ozone was still below the 1964–1980 average (Executive Summary). - Recovery is noisy from year to year. - The 2020–2023 Antarctic holes were deep and long-lasting. - WMO’s Ozone and UV Bulletin says none of the short-term influences behind them is linked to the decline in ozone-depleting substances. They therefore “do not undermine confidence” in recovery. - The 2024 hole was below the 1990–2020 average. Its late onset is described as “a robust indication of initial recovery” (WMO Ozone and UV Bulletin No. 3, Sept 2025). - The 2025 hole was the fifth smallest since 1992. NASA and NOAA say that since peaking around 2000, Antarctic stratospheric ODS levels “have declined by about a third”, measured relative to pre-hole levels. This is consistent with NOAA’s index being 28% of the way back for Antarctica. They project full recovery “around the late 2060s”, and name legacy stocks in “old products like building insulation and in landfills” as a reason recovery is slow (NASA Science, 24 Nov 2025). - The negotiating history. Two WMO sources date the turning point earlier than the chapter does. - WMO’s 1994 assessment was written when the 1992 Copenhagen provisions were the latest. It projected that the Antarctic ozone hole would recover “in about the year 2045, other things being equal”. It also said recovery “would have been impossible without” the amendments to the 1987 Protocol (WMO 1994 Executive Summary). - WMO’s 2022 retrospective. Q14 of the 2022 Twenty Questions says the 1987 and 1990 provisions only slowed EESC growth. It dates the first projected long-term decrease to the 1992 Copenhagen Amendment (Twenty Questions and Answers About the Ozone Layer: 2022 Update, Q14). - Farman’s claims about 1987 and 1990 are consistent with this. His claim that only the 1995 adjustments secured a return is not, at least on WMO’s framing. - His source (DETR, 1999) may have assumed continued, uncontrolled growth in developing-country HCFC and methyl bromide use under the Copenhagen provisions. That could explain the difference, but I could not retrieve DETR (1999) to check. - The ratchet kept moving after 2001: the 2007 HCFC acceleration (Decision XIX/6) and the 2016 Kigali Amendment (see Claim 5). This supports the chapter’s picture of stepwise tightening at consensus pace. - Minor technical update. WMO 2022 puts bromine at about 60–65 times as effective as chlorine per atom, against the chapter’s 58 (WMO/UNEP 2022, Ch. 1).

Verdict: partly held up.

• Held up. The direction and order of magnitude of the forecast (a return after mid-century, not before) held, as did the diagnosis that early agreements were inadequate.
• Too early for Antarctica. The 2050–2060 window is too early for the Antarctic and for WMO’s current mid-latitude metric.
• Not supported. The claim that only the 1995 adjustments secured a projected return is not supported by WMO’s own contemporaneous (1994) and retrospective (2022) accounts.

Implication for weight.

• The persistence lesson is strengthened. Since 2001, projected recovery dates have repeatedly moved later as hidden stocks and leaks came to light.
• Treat forecast dates cautiously. Any lesson built on the chapter’s specific dates, or on its attribution of the turning point to 1995, should carry little weight.

Claim 2: CFC-12 will persist past 2100, fed by banks#

Original claim (p. 77).

• Persistence. CFC-12 has a lifetime of about 100 years, so in 2100 its atmospheric concentration will be “at least 37%” of its 2001 level. The 37% is simply e⁻¹ over one lifetime.
• Why “at least”. Even where production has ceased, releases continue from equipment and foams, and this bank was estimated at 791 kt in 1995.
• Comparison. Methyl chloroform (lifetime about 5 years) will be down to about 2% after 20 years.

Subsequent developments.

• Lifetimes barely changed. WMO 2022 gives CFC-12 a lifetime of 102 years (±15%) and methyl chloroform 5.0 years (WMO/UNEP 2022, Ch. 7, Table 7-1).
• The observed and projected trajectory (WMO/UNEP 2022, Ch. 7, Appendix Table 7A-1):
• CFC-12 was 542.9 ppt at the start of 2001;
• it peaked at 543.6 ppt in 2002–2003;
• it was 500.8 ppt in 2020;
• the baseline projects 236.3 ppt in 2100, about 44% of the 2001 value.

The “at least” qualifier was therefore warranted. - Recent emissions. CFC-12 fell about 2.8% over 2016–2020, with global emissions of 25 ± 20 Gg per year in 2020. Emissions from north-eastern China fell from 3.3 to 0.5 Gg per year between 2016 and 2019, consistent with co-production alongside illicit CFC-11 (WMO/UNEP 2022, Ch. 1). - The banks are larger than earlier assessments assumed. - A Bayesian analysis found CFC-11 and CFC-12 banks “larger than recent international scientific assessments suggested”. Left unrecovered, they could delay Antarctic recovery by about six years and add about 9 billion tonnes of CO2-equivalent emissions (Lickley et al., Nature Communications, 17 Mar 2020). - WMO 2022 adopted this approach. It found that future releases from CFC banks contribute more to EESC than releases from HCFC or halon banks, and that eliminating them would bring the mid-latitude return forward by about 5 years. It also notes that “estimates of bank sizes are highly uncertain” and that it made no judgement about how accessible banks are for capture (Ch. 7). - Policy on banks. - There is still no global obligation to recover and destroy CFC banks. - The EU’s 2024 ODS Regulation requires that emissions from ODS-containing insulation foams in old buildings be avoided during renovation or demolition, with the ODS ultimately destroyed (European Commission, Regulation (EU) 2024/590 summary, accessed Sept 2026; Regulation (EU) 2024/590, EUR-Lex). - Methyl chloroform, as a check. It fell from 41.5 ppt in 2001 to about 1.4 ppt in 2020, “1% of its maximum value” (Ch. 1; Ch. 7, Table 7A-1).

Verdict: held up. The persistence arithmetic was correct. The qualification about banks was, if anything, understated.

Implication for weight. The digest’s insight 11 (ending production does not end harm, because legacy stocks remain) can carry strong weight. The post-2001 record adds that the size of legacy stocks is itself uncertain, and was underestimated by the institutions responsible.

Claim 3: The long-lived halon will keep rising until about 2020 unless banks are destroyed#

Original claim (p. 77).

• The substance. The text does not name the halon. The figures (a lifetime of about 65 years, and developed-country production ending in 1993) point to halon-1301.
• Production and banks. Production peaked in 1988 and ceased in developed countries at the end of 1993. The developed-country bank was estimated at 70 kt in 1995 and was being run down at about 4% a year.
• The forecast. Atmospheric concentration will continue to rise “until about 2020, unless the protocol is amended to require the destruction of banked halons”.

Subsequent developments.

• The trajectory matched the forecast.
• WMO 2022 found “no significant change” in halon-1301 between 2016 and 2020. At about 3.3 ppt (3.37 ppt by AGAGE, 3.32 by NOAA) it is now the most abundant halon.
• Emissions were stable at about 1.3 Gg per year over 2016–2020 (WMO/UNEP 2022, Ch. 1, §1.2.2).
• The WMO baseline series rises from 2.86 ppt (2000) to 3.34 ppt (2015), stays at about 3.35 ppt through the 2020s, then declines slowly to about 2.24 ppt in 2100 (Ch. 7, Table 7A-1).
• The revised lifetime is 72 years (±13%).
• Destruction was never mandated, and policy moved the other way.
• Decision XXX/7 (2018). The Parties noted “with concern” that civil aviation could face a lack of halons “in the upcoming decades” to service aircraft being built now. They asked for ways to recover more halon from ship-breaking (Decision XXX/7, Nov 2018).
• The Parties’ technical panel (TEAP) reports that all aircraft still depend on halon from stocks for most fire protection. It estimates halon-1301 will run out for civil aviation between 2030 and 2049. It warns that destroying halon-1301 for carbon credits could create shortages, and that halon had been destroyed through “lack of understanding about long-term needs” (TEAP 2022 Assessment Report, pp. 9–10, 16, pub. Apr 2023).
• The EU. Its 2024 ODS Regulation provides that recovered halons should not be destroyed unless they cannot be recycled or reclaimed for continued fire protection, “particularly aboard aircraft” (European Commission ODS legislation page).
• Production continues for one feedstock use. WMO 2022 lists halon-1301 among ODSs still produced as a feedstock (for the insecticide fipronil), with estimated emission rates of a few per cent (Ch. 7, Table 7-1 and §7.2.2).

Verdict: held up. The forecast was accurate: the rise continued until the mid-2010s and was followed by a plateau around 2020. The chapter did not anticipate why its remedy would be rejected: without drop-in alternatives for some uses, the banked stock became the only supply.

Implication for weight.

• The forecasting skill is well demonstrated and supports the chapter’s credibility on persistence.
• The halon story adds a caution to insight 11. Legacy stocks can become a needed resource that locks in continued use, so “destroy the bank” is not always the right answer. This tension is absent from the chapter, which treats destruction as simply unfinished business (pp. 77, 81).

Claim 4: Additional skin cancer, peaking around 2050 (Fig. 7.1)#

Original claim (p. 76, Fig. 7.1 and caption).

• Harms. Ozone depletion will cause additional skin cancer, as well as cataracts and immune suppression.
• Timing. Given a 30–40-year latency, the schematic (after Slaper et al., 1996) shows the excess starting around 2010, peaking around 2050 and declining by the 2080s.
• The caption’s caveats. The caption says reality is “far more complex”. It notes a rising background rate of skin cancer and the role of human behaviour.

Subsequent developments.

• The source study. Slaper et al. modelled the Copenhagen scenario: an ozone minimum around 2000, then a peak relative increase in skin cancer incidence of “almost 10%” about 60 years later (Slaper et al., Nature, Nov 1996).
• Later modelling by the same group supports the shape and shrinks the size. van Dijk et al. (Slaper a co-author) coupled chemistry-climate models to a UV risk model. They found:
• under full compliance, excess incidence “will peak mid 21st century and then recover or even super-recover”;
• the peak-year excess is “up to 4 %” of total local incidence;
• this is 30–40 extra cases per million per year in Western Europe and 170–200 in north-east Australia;
• the Protocol would prevent about 2 million skin cancer cases a year by 2030 (van Dijk et al., Photochemistry and Photobiology, online 26 Sept 2012, 2013 issue).
• WMO 2022 says the projected additional cases from ozone depletion are largest “in the first half of the 21st century” (Twenty Questions 2022, Q16).
• Estimates of avoided harm have grown and been formalised.
• US EPA estimates that full implementation of the Protocol will prevent about 443 million skin cancer cases, 2.3 million skin cancer deaths and 63 million cataract cases among Americans born 1890–2100 (US EPA, “Health and Environmental Effects of Ozone Layer Depletion”, citing its May 2020 AHEF report).
• The Protocol’s Environmental Effects Assessment Panel (EEAP) cites 11 million melanomas, 432 million keratinocyte cancers and 63 million cataracts prevented in the same US cohort (Neale et al., EEAP, Photochem. Photobiol. Sci., 1 Mar 2023).
• The 2013 Late lessons volume (Ch. 23, Box 23.3, p. 576) cites about 47,000 skin cancer cases a year avoided in north-western Europe by 2050. It also cites 14,000 extra cases a year still expected from damage already done.
• Observed trends are dominated by other factors.
• EEAP reports that skin cancer incidence “continues to rise” but has stabilised in younger populations in some countries. It also reports new evidence that UV exposure affects the immune system in both harmful and beneficial ways (Neale et al., 2023).
• None of the assessments I checked reports the ozone-attributable excess as detected in cancer registries. It remains a modelled quantity.

Verdict: partly held up.

• Held up. The qualitative claims (latency-driven delay, a mid-century peak, cataracts and immune effects) are supported by later modelling and assessment.
• Not tested. The schematic’s timing has not been tested against observation and probably cannot be.
• Smaller than the source implied. Later work puts the magnitude well below Slaper’s “almost 10%”.
• More complex than the caption said. The immune claim now has a benefit side.

Implication for weight. The latency lesson (harm lags cause by decades, so the damage is committed well before it is seen) remains sound, but it is carried by models, not by observed health data. The huge “avoided” numbers describe the world without the Protocol. They should not be read as the harm the chapter’s timeline implies.

Claim 5: The open-ended HCFC/HFC transition was “deeply flawed”#

Original claim (p. 81, fn 7; also p. 76).

• The decision. In 1990 the negotiators accepted the industry’s preferred substitutes (HCFCs and HFCs). They made them subject to “guidelines rather than controls” and left their future open-ended.
• Farman’s judgement. This was “deeply flawed”, for three reasons:
• HFCs are powerful greenhouse gases;
• the Multilateral Fund (MFMP) was spent largely on replacing CFCs with HCFCs, so “more money is now needed” to phase out HCFCs;
• radical halocarbon-free, energy-efficient technology should have been encouraged from the outset.

Subsequent developments.

• The Parties accelerated the HCFC phase-out and paid twice (Decision XIX/6, Sept 2007).
• Timetable. Developing (Article 5) countries: a freeze in 2013, then 10% (2015), 35% (2020) and 67.5% (2025) reductions, and phase-out by 2030 with a 2.5% servicing tail to 2040. Developed countries: phase-out by 2020.
• Funding rules. The same decision told the Multilateral Fund’s Executive Committee to change its eligibility rules on “post-1995 facilities and second conversions”, meaning paying again for enterprises it had already converted to HCFCs. This is exactly the double cost Farman predicted.
• Choice of substitutes. It urged Parties to choose substitutes that minimise environmental impacts, “in particular impacts on climate”.
• Follow-through. HCFC phase-out management plans were approved from 2010 onward (Multilateral Fund, history timeline).
• Scale. By September 2026 the Fund reported USD 4.3 billion in grants across 10,198 projects in 144 countries (Multilateral Fund homepage, accessed 25 Sept 2026). I could not break out the HCFC share.
• HFCs were brought under control.
• The warning. Velders et al. warned that unregulated HFC growth, driven by the HCFC phase-out, could reach 9–19% of projected global CO2 emissions by 2050 (PNAS, 22 June 2009).
• The response. The Kigali Amendment (adopted 2016, in force 2019) set an HFC phase-down. It had 174 parties at 10 August 2026. The US ratified on 31 October 2022 (Ozone Secretariat ratification status).
• The benefit. WMO 2022 estimates that compliance with Kigali avoids 0.3–0.5 °C of warming by 2100 (Executive Summary).
• The EU now plans to phase HFCs out entirely by 2050 (F-gas Regulation (EU) 2024/573, adopted 7 February 2024) (European Commission).
• The US continues to allocate HFC allowances under the AIM Act (the 2020 US law implementing the HFC phase-down). By September 2026 it had also finalised a rule reconsidering some technology-transition requirements and proposed exempting transport refrigeration units from leak-repair rules; the page does not date these actions (US EPA, accessed Sept 2026).
• Second-round problems from the substitutes’ chemistry.
• HFC-23. This by-product of HCFC-22 production reached record emissions in 2018, although reported abatement should have cut them by 87% (Stanley et al., Nature Communications, 21 Jan 2020).
• CFC by-products. Five CFCs rose over 2010–2020, three of them “probably” as by-products of HFC production (Western et al., Nature Geoscience, Apr 2023).

• TFA. Hydrofluoroolefin (HFO) replacements for HFCs degrade to trifluoroacetic acid (TFA), a very persistent breakdown product. Over Europe, HFO-1234yf may already dominate atmospheric TFA generation (Holland et al., ES&T Letters, 15 June 2026).

  The Protocol’s Environmental Effects Assessment Panel judges the risk from TFA “currently de minimis” (EEAP Update 2024, 17 Mar 2025). TEAP notes that some definitions of PFAS (the “forever chemicals” now facing broad restriction) include most in-kind halon alternatives, so restricting PFAS could restrict them (TEAP 2022, p. 16). - Halocarbon-free technology proved viable at scale in some sectors. - Around 200 million new domestic refrigerators and freezers are sold each year, “dominated by the HC-600a refrigerant technology” (isobutane). - Consumer aerosols have largely moved to hydrocarbons and dimethyl ether (TEAP 2022, pp. 18, 20, 90). - Counter-evidence and limits. - The HCFC detour’s ozone cost was modest and is being unwound. HCFCs made up about 10% of tropospheric chlorine in 2020 (WMO 2022, Ch. 1). Their equivalent effective chlorine and radiative forcing peaked in 2021, five years earlier than projected (Western et al., Nature Climate Change, 11 June 2024). - The counterfactual remains untested. Farman held that a halocarbon-free route from 1990 would have been better. The opposing view is that HCFCs as drop-in transitional chemicals made the fast CFC exit feasible. Neither has been tested. Later hydrocarbon dominance in domestic refrigeration shows feasibility in that sector, not that a faster route existed in 1990 across all sectors.

Verdict: strengthened.

• Strengthened. The Parties themselves adopted the chapter’s diagnosis: they accelerated the HCFC phase-out, paid for second conversions and controlled HFCs.
• Still contested. The counterfactual (that a radical halocarbon-free route was available and better in 1990) remains contested.

Implication for weight. Insight 10 (substitution shaped by incumbent producers seeds second-round problems) can carry strong weight. The post-2001 record extends it:

• each substitute generation (HCFC, then HFC, then HFO) created a new problem;
• the new problems were climate forcing, by-product emissions and a persistent breakdown product;
• each correction required new controls and new money.

Claim 6: Developing-country phase-out by 2010, bought out by the Multilateral Fund#

Original claim (pp. 77, 81).

• Supply to developing countries. Their phase-out began with a 1999 freeze, and supply was to cease in 2010.
• Buy-outs. In March 1999 the Fund approved USD 150 million to close China’s CFC production over 10 years, and later that year USD 82 million for India.
• Halons and bromochloromethane. Production of the halon in China, India and Korea was “due to cease in 2002”. A phase-out of bromochloromethane (BCM) production by 2002 was agreed at Beijing.

Subsequent developments.

• Delivered on schedule. The Fund’s timeline records:
• the 1999 freeze at 1995–97 levels;
• 50% cuts in CFCs and halons by 2005;
• an 85% cut in CFCs by 2007;
• “total elimination” of CFCs, halons and carbon tetrachloride in developing countries by 2010;
• complete phase-out of BCM in 2002 (Multilateral Fund history).
• India. The Fund’s Executive Committee approved USD 82 million at its 29th meeting (November 1999). India completed its CFC production phase-out on 1 August 2008, 17 months ahead of schedule. After 2008, production continued only for metered-dose inhalers under essential-use exemptions. The World Bank’s completion report also records more customs confiscations in 2004–2008 (World Bank Implementation Completion and Results Report ICR00002241, 22 June 2012).
• China.
• In January 2007 the World Bank reported that China, India and Argentina had committed to stop CFC production in 2007, ahead of the 2010 deadline (World Bank, Montreal Protocol 2007 Business Plan, 29 Jan 2007).
• On 1 July 2007 China shut five of its six remaining CFC and halon plants, “two and a half years ahead” of the 2010 deadline. This cut CFC production to about 550 tonnes, all for medical inhalers, from a peak of 55,000 tonnes in 1998. Thirty-one other CFC and halon plants had already closed with Multilateral Fund and World Bank support (World Bank press release, 1 July 2007, archived copy).
• I still found no primary document for the USD 150 million total approved in March 1999.
• The halon date is not supported for China and does not match the treaty schedule.
• For developing countries, 2002 was the halon freeze year: a 50% cut was due in 2005 and phase-out in 2010. That is consistent with the Fund’s timeline above.
• China’s own plan. In November 2000 China’s plan was to end use of halon-1211 by the end of 2005 and of halon-1301 by 1 January 2010 (People’s Daily, 23 Nov 2000; contemporaneous news report).
• When production ended. The World Bank’s July 2007 release says China had “recently ended the production of halon for emissive use”. That places the end of Chinese halon production for fire protection around 2006–2007, not 2002.
• India and Korea. I found no evidence either way. The text may have confused the 2002 freeze with cessation.
• Halon-1301 is still produced as a feedstock (Claim 3).
• Legal phase-out did not end production. After 2012, CFC-11 was produced without being reported in eastern China, in breach of the 2010 phase-out (Claim 7).

Verdict: held up. The CFC phase-out and the buy-outs happened as described. Both countries finished early: India about 17 months early, and China two and a half years early for all but inhaler-grade CFCs. The one error is the 2002 halon date. It conflicts with the treaty schedule, and for China it is contradicted by the 2007 World Bank account.

Implication for weight. Paying producers to close capacity worked. The follow-on lesson is that a legal end date is only as good as the verification behind it (see Claim 7).

Claim 7: Illegal trade and the unresolved business of 1999#

Original claim (p. 81; Table 7.1).

• The threat. The Beijing Declaration (1999) called for continued efforts to halt illegal trade in ODS.
• Unresolved issues:
• destroying halon banks;
• an earlier phase-out of methyl bromide;
• CFC and halon consumption in Russia and other economies in transition.
• Precaution. There was “one brief glimpse” of precaution: a request for ways to prevent new ozone-depleting chemicals being developed and marketed.

Subsequent developments.

• Illegal production turned out to be a bigger threat than illegal trade.
• Detection. Montzka et al. showed that the decline in atmospheric CFC-11 slowed by about 50% after 2012. This implied emissions about 13 Gg per year higher, “despite reported production being close to zero” and suggesting “unreported new production” (Nature, 16 May 2018).
• Attribution. Rigby et al. traced at least 40–60% of the rise to eastern mainland China, mainly Shandong and Hebei provinces (Nature, 22 May 2019).
• Response. The Parties responded with Decision XXX/3 (November 2018), expressing “serious concern” and commissioning reviews of monitoring, reporting and verification (Decision XXX/3).
• Reversal. By 2019 global emissions had fallen by 18 ± 6 Gg per year (Montzka et al., Nature, 10 Feb 2021), and eastern China’s CFC-11 emissions had returned to pre-2013 levels (Park et al., Nature, 10 Feb 2021).
• Lasting cost. The CFC-11 bank may have grown by 90–725 Gg, and WMO 2022 puts the delay to mid-latitude recovery at about one year (Ch. 7).
• How it was found. The problem was detected by long-term remote monitoring stations, including the Gosan (South Korea) and Hateruma (Japan) stations. This vindicates the chapter’s point about the value of systematic monitoring (p. 82).
• Carbon tetrachloride remained unexplained for years.
• Its abundance continued to fall more slowly than expected. WMO 2022 puts global emissions at 44 ± 14 Gg per year in 2020 and says there are still indications of a gap between reported and observed emissions (WMO 2022, Ch. 1, §1.2.3).
• Emissions in eastern China rose after 2013 to 11.3 kt per year in 2016, then fell to 6.3 kt per year in 2019, tracking CFC-11 production.
• The SPARC report on the “mystery” of carbon tetrachloride (2016) narrowed the gap through revised lifetimes and newly identified industrial sources. It is cited by WMO 2022 (SPARC Report No. 7), but I could not open it.
• Feedstock exemptions became a significant leak. See Claim 1 (Reimann et al., 2026). Proposals to narrow the exemptions had been made earlier (Andersen et al., PNAS, Dec 2021).
• 2024. The Parties’ response in Decision XXXVI/5 asks Parties to “take steps to minimize emissions” during production and feedstock use, and invites voluntary reporting of national procedures. It leaves the Article 1(5) exemption unchanged (Decision XXXVI/5, MOP-36, 2024).
• 2025. At MOP-37 (Nairobi, 3–7 November 2025) the Scientific Assessment Panel flagged “increased production of ODSs for feedstock uses”, and the meeting again reached no agreement on feedstocks.
• Still open. The Secretariat’s summary also records that HFC-23 shows “significant gaps” between reported and measured emissions, “with no new sources identified since 2015”. The Panels are to report again at MOP-38 in November 2026 (Ozone Secretariat, MOP-37 meeting summary, accessed 25 Sept 2026; IISD Earth Negotiations Bulletin, MOP-37 summary report, Nov 2025, secondary, not opened in full).
• Russia and the economies in transition were resolved quickly.
• Decision XIII/17 (2001) records that Russia had been in non-compliance for 1999–2000. It records that Russia closed CFC production on 20 December 2000 and stopped Annex A and B trade from 1 March 2000, with help from a donor-funded special initiative (Decision XIII/17).
• Decision XIV/35 (2002) confirmed complete phase-out of Annex A and B production and consumption in 2001 (Decision XIV/35).
• Methyl bromide was mostly phased out, apart from one large exemption.
• Controlled uses ended in 2005 in developed countries and 2015 in developing countries. Critical-use exemptions have since fallen about 200-fold.
• Quarantine and pre-shipment (QPS) use, which the Protocol does not control, now accounts for almost all reported production: about 99% (WMO 2022, Ch. 7) and 100% in 2021 (TEAP 2022). QPS use runs at about 10,000 tonnes a year.
• Ending QPS emissions would bring recovery forward by about two years (WMO 2022, Ch. 7; TEAP 2022, p. 10).
• Halon banks. They were not destroyed and are now deliberately conserved (Claim 3). MOP-37 again discussed the “continued use of halon 1301 in aviation” and uncertainty about long-term supply, and invited information on alternatives for a TEAP report due in 2027 (Ozone Secretariat, MOP-37 summary).
• Unwanted refrigerant banks are back on the agenda. According to the ENB summary (seen only as a search excerpt), more than 100 developing countries have inventoried their unwanted ODS banks. MOP-37 asked TEAP to estimate the global quantity of used and unwanted refrigerants and the benefits of reclaiming or destroying them. This builds on the 2024 decision on life-cycle refrigerant management (Decision XXXVI/2; ENB summary, above). Twenty-five years after Beijing, destroying banks is being studied again, but for refrigerants rather than halons.
• Unregulated gases the chapter did not discuss.
• Nitrous oxide is now “the single most important ozone-depleting emission” and is not regulated by the Protocol (Ravishankara et al., Science, 27 Aug 2009).
• Dichloromethane, a very short-lived chlorinated solvent, more than doubled in abundance since the early 2000s (WMO 2022, Ch. 1).

Verdict: strengthened.

• The threat materialised. It was more serious than the chapter’s warning implied: it came from covert production rather than trade.
• The regime proved able to detect it and reverse it.
• Unresolved items. Russia was resolved within a year or so. Methyl bromide was resolved except for the QPS exemption. Halon banks went the opposite way to the chapter’s prescription.
• New unresolved items. A new set replaced them: feedstock emissions, unexplained HFC-23 and carbon tetrachloride emissions, and unwanted refrigerant banks. As of MOP-37 (November 2025) the Parties had adopted no new binding controls in response to any of them. They were relying on studies, reporting and calls to minimise emissions.

Implication for weight. The lesson that compliance declared on paper can hide non-compliance, and that independent verification matters, can carry strong weight. So can the monitoring lesson on p. 82. The CFC-11 episode is also evidence for the regime’s adaptive capacity. That tempers any reading of the chapter as showing that global regimes simply fail.

Claim 8: 1977–80 was the “first, and last” precaution; 1987 was “overtaken by events”#

Original claim (p. 80).

• The only precautionary step. The 1977–80 aerosol measures (in the US, Canada, Norway and Sweden, and the EC’s Decision 80/372) were “the first, and last, unequivocal application of the precautionary principle in the ozone story”. They were the only action taken before the evidence became compelling.
• 1987. Benedick’s claim that the 1987 Protocol was precautionary is rejected. The working group had been “overtaken by events” (the 1985 Antarctic findings), and the timing of the signature was “a pre-emptive move” ahead of the 1987 Antarctic expedition results.
• The ultimate objective. On p. 81 Farman adds that the Protocol’s “ultimate objective was not clearly defined”.

Subsequent developments.

• The treaty’s own words cut both ways (Montreal Protocol, preamble, consolidated text; Vienna Convention, preamble).
• For Farman. The preamble notes “the precautionary measures for controlling emissions of certain chlorofluorocarbons that have already been taken at national and regional levels”. The Vienna Convention (1985) similarly notes precautionary measures “already taken”.
• Against him. The preamble also declares the Parties “determined to protect the ozone layer by taking precautionary measures”, with “the ultimate objective of their elimination”. That also qualifies his remark that the ultimate objective was undefined, although the 1987 control measures did not implement elimination.
• Original wording confirmed. Both paragraphs (“NOTING the precautionary measures…” and “DETERMINED to protect the ozone layer by taking precautionary measures…”) appear in the original 1987 text as published in the EC’s Official Journal (Montreal Protocol, OJ L 297/21, 31 Oct 1988, EUR-Lex). The precautionary self-description dates from the signing, not from later amendments.
• Benedick’s 2005 testimony, the primary statement the digest refers to, was given to the US Senate Committee on Environment and Public Works on 28 September 2005, under the title “The Case of the Montreal Protocol: Science Serving Public Policy” (testimony PDF). On the points in dispute:
• The evidence at the time. During the negotiations “the arguments for controlling CFCs rested on unproven scientific theories”. Thirty years of measurements had shown no statistically meaningful mid-latitude depletion. Most scientists then considered the Antarctic hole “an anomaly”, since it did not fit the models and might have had non-human causes.
• The decisive evidence came after signing. Preliminary results from the 1987 Antarctic expedition were announced “about two weeks after the protocol’s signing”, and the Ozone Trends Panel report followed in March 1988. That confirms the chapter’s own chronology (p. 80). Benedick reads it as proof that the signatories acted before the evidence was in.
• His characterisation. The treaty was “preventive action on a global scale” against “speculative future dangers”.
• Common ground with Farman. Benedick credits the 1977 Clean Air Act’s “may reasonably be anticipated” standard as a deliberately low threshold for intervention, which supports Farman’s view that the 1977–80 measures were precautionary. He also says the Montreal models “had proven incapable of predicting” the Antarctic losses.
• What the testimony does not address. It does not engage Farman’s argument that the Antarctic findings, not the theory, changed the negotiating politics in 1985–87. It is a participant’s account. It was given at a hearing, “The Role of Science in Environmental Policy Making” (S. Hrg. 109-1003), that dealt largely with contested climate science (GovInfo hearing record).
• Benedick’s other statements. He had set out the same position in the enlarged 1998 edition of Ozone Diplomacy and did so again in a 2009 chapter (Benedick, Ozone Diplomacy, enlarged ed., Harvard UP, 1998; Benedick, “Science Inspiring Diplomacy: The Improbable Montreal Protocol”, in Twenty Years of Ozone Decline, Springer, 2009). I could not open either text.
• Institutional memory now leans towards Benedick. The Multilateral Fund’s history page calls the Vienna Convention “the first major global treaty employing the precautionary principle” (Multilateral Fund history). The 2013 Late lessons volume calls the ozone case “a cause célèbre” for international agreements (Ch. 23, Box 23.3, p. 575). Neither engages Farman’s argument.
• Scholarship. Parson (2003) gives the most detailed reconstruction of the 1980–87 negotiations and of industry strategy (Parson, Protecting the Ozone Layer: Science and Strategy, OUP, 2003). I did not re-read it for this check. Before the synthesis relies on either side, his account of how the Antarctic findings affected the 1987 negotiations should be checked.

Verdict: contested.

• Largely definitional. On a formal definition (acting before proof of damage), 1987 qualifies as precautionary. On a political-economy reading (what actually moved the negotiators), Farman’s account of a shock-driven, credibility-saving agreement remains plausible, but primary sources do not settle it.
• Benedick’s 2005 testimony sharpens the dispute without resolving it. The two participants agree on the chronology and on the models’ failure. They disagree on whether an anomaly that most scientists had not yet attributed to CFCs moved the negotiators. Each speaks from his own role in the story (Farman as the discoverer of the losses, Benedick as the US negotiator).
• Clearly wrong. “Last” is too strong. The treaty’s own framing and later practice both use precautionary language, including the 1999 Beijing request that Farman himself notes, and the 2007 and 2016 decisions taken before full harm had materialised.

Implication for weight.

• Timing lesson: robust. The digest’s insight 6 holds for timing: a plausible mechanism brought limited action, and dramatic confirmation brought decisive action.
• Classification: weak. The claim that 1987 was not precautionary should carry little weight on its own.
• A sharper formulation. The ozone case shows precaution in the formal sense (acting before attribution) happening because a vivid, unexplained anomaly had shifted the politics.

Claim 9: DuPont’s pledge, its denial until 1986, and the first-mover penalty#

Original claim (p. 80).

• The pledge. In a full-page New York Times advertisement on 30 June 1975, DuPont pledged to stop production if “reputable evidence” showed harm.
• The reversal. It denied such evidence existed until 1986. In September 1986 it wrote to customers accepting the need for some controls (citing Cagin and Dray, 1993).
• Industry research funding. Industry did fund substantial research through the Chemical Manufacturers Association.
• The first-mover penalty. The US refused further unilateral action. Its share of world CFC-11/12 production had fallen from 46% (1974) to 28% (1985).

Subsequent developments.

• Industry strategy. Later business-strategy work treats the 1986 shift partly as commercial positioning. DuPont had a lead in developing substitutes and could profit from a regulated transition (Maxwell & Briscoe, “There’s money in the air: the CFC ban and DuPont’s regulatory strategy”, Business Strategy and the Environment, Nov 1997).
• This is consistent with Farman’s own account on p. 81: the 13 company groups were content to close CFC plant if given time to profit from HCFC and HFC investment.
• I could not access the abstract; the characterisation rests on the article’s title and standard citation.
• Industry’s September 1986 shift was coalition-wide. Benedick’s 2005 testimony records that the Alliance for Responsible CFC Policy, a coalition of about 500 producer and user companies, “announced its acceptance of international controls in September 1986” (Benedick testimony, 28 Sept 2005). That corroborates the chapter’s September 1986 date for DuPont’s change of position.
• Industry funding of research. Benedick also corroborates the chapter’s point that the Chemical Manufacturers Association funded ozone research. He adds that when the UK government stopped funding the British Antarctic mission in 1985, the CMA filled the gap. I have not verified this from a second source.
• DuPont’s 1988 commitment. In March 1988, days after the Ozone Trends Panel report, DuPont announced that it would go beyond the Protocol and end CFC production entirely, targeting 2003 (UPI, 4 Sept 1988, a contemporaneous news report; C&EN, “CFC production: Du Pont seeks total phaseout”, Apr 1988, not opened).
• In effect DuPont honoured the 1975 pledge only after global loss had been formally attributed.
• The chapter omits this step (see the digest’s omissions).
• The 2013 volume adds a link. It connects the same chemist (Midgley) and the GM–DuPont relationship to tetraethyl lead (Late lessons II, Ch. 3, p. 50).
• The 46% → 28% figures were not re-checked. I found no later source that contradicts them.

Verdict: held up. The dated facts are consistent with later accounts. The main refinement is that the 1986 shift reflected commercial readiness as well as evidence. That supports, rather than weakens, the chapter’s broader argument about incumbents shaping the transition.

Implication for weight. Insight 5 (whoever sets the evidential threshold sets the timing of action) is well illustrated but rests largely on this one case. Insight 8 (the first-mover penalty) rests on one statistic that was not re-checked, so it should stay at moderate weight.

Claim 10: No mandate for global experiments; technology outstrips the science#

Original claim (pp. 82–83).

• The context. The chapter is framed against the US decision in 2001 not to ratify Kyoto.
• No mandate. Neither governments nor multinational companies have a mandate for global experiments, “even when such experiments consist solely of ‘business as usual’”.
• Technology ahead of science. “All too often technology outstrips the science needed to assess the risks”. Policy-makers must recognise much faster when “ignorance has been replaced by understanding, however rudimentary”.
• The inertness paradox. Short-term safety demands inert chemicals, but inert means persistent.

Subsequent developments.

• The contrast between the two regimes widened.
• Kyoto entered into force on 16 February 2005 without the US, which signed in 1998 but never ratified (UN Treaty Collection, Kyoto Protocol status, accessed 25 Sept 2026).
• The US left the Paris Agreement in 2020, rejoined in 2021, and notified withdrawal again on 27 January 2025, effective 27 January 2026 (UN Treaty Collection, Paris Agreement status, note 7; Executive Order, 20 Jan 2025).
• A presidential memorandum of 7 January 2026 directed withdrawal from the UNFCCC and the IPCC (White House, 7 Jan 2026). The Montreal Protocol and Vienna Convention are not on the list.
• Meanwhile the ozone treaties have 198 parties and Kigali 174, including the US (Ozone Secretariat).
• So the hope that the Montreal experience would carry over to climate was, as Farman feared, largely disappointed. The Montreal regime itself proved more durable, and it took on part of the climate task through Kigali.
• “Technology outstrips the science” was repeatedly confirmed within the ozone regime. Each of these was a surprise relative to the assumptions built into controls:
• unreported CFC-11 production (Claim 7);
• HFC-23 emissions contrary to reports (Claim 5);
• CFCs appearing as by-products of HFC manufacture (Claim 5);
• feedstock losses about seven times the assumed rate (Claim 1);
• TFA from the newest substitutes (Claim 5);
• banks larger than assessed (Claim 2).

The regime’s own science panels caught most of these, sometimes a decade after the activity began. - Monitoring caught what inventories missed. Surface networks detected CFC-11, and satellites plus models allowed formal detection of recovery (Wang et al., 2025). This bears out the chapter’s point that systematic long-term measurement is what reveals the unexpected (p. 82). - WMO’s 2025 Bulletin makes the same point in its introduction: there remains “an essential need” for “careful systematic monitoring”. It also reports that ageing ground-based instruments are no longer manufactured (WMO Ozone and UV Bulletin No. 3, Sept 2025). The chapter’s complaint that long-term monitoring is hard to fund (p. 82) remains current. - The satellite story. On the satellite data, Farman’s measured wording has held up better than later retellings. - Farman’s wording (p. 82). NASA scientists were “re-examining” low values “flagged as ‘suspect’” by their software. - Later retellings. Benedick’s 2005 testimony says the satellite computers were programmed “to automatically reject” such values, and the 2001 report’s editors say they were “set aside”. - Historians’ accounts. They say the values were flagged, not discarded, and were checked against ground data that seemed to show normal levels (Christie, The Ozone Layer: A Philosophy of Science Perspective, CUP, 2001; Pukelsheim, IMS Bulletin, 1990; both as summarised in Hyndman, “How NASA didn’t discover the hole in the ozone layer”, 29 Sept 2023, a secondary blog source. I did not open the originals). - The framing was taken up. - “Stratospheric ozone depletion” is one of the nine planetary boundaries in the 2009 framework (Rockström et al., Nature, Sept 2009). - A 2026 perspective traces an “atmospheric track” of chemicals of global concern, exemplified by CFCs, alongside the persistent-pollutant track. It asks why chemical regulation “repeatedly failed to prevent widespread contamination” and proposes class-based phase-outs and pre-market controls (Scheringer, Arp & Cousins, Environ. Sci. Technol., 2 Feb 2026). - Both echo Farman’s “global experiment” and inertness–persistence arguments, although neither cites him directly as far as the abstracts show.

Verdict: strengthened.

• Descriptive claims: strengthened. Technology running ahead of assessment, and inert meaning persistent, are borne out by the post-2001 record.
• Normative claim: not testable. “No mandate” is a value claim and cannot be tested.
• The climate contrast: borne out. It developed much as Farman’s pessimism implied.

Implication for weight.

• Insights 12 and 13 in the digest remain assertions. These are: recognise early when ignorance has become understanding; and there is no mandate for global irreversible experiments. The hindsight record makes them more plausible, but does not prove them.
• Insight 1 can be upgraded. This is the paradox that safety at the point of use creates hazard at system scale. It now has support from a second substitution cycle: HFOs chosen to be short-lived yield a very persistent product.

What the hindsight check changes about the section’s lessons#

This table maps the digest’s transferable insights (technology-neutral) to their strength after this check. Page references are to the 2001 chapter.

Lessons the hindsight record adds, stated in technology-neutral terms:

1. Exemptions built on unverified low-loss assumptions become the next leak. Carve-outs justified by expected minimal emissions can grow into significant sources when the assumption is never checked against independent measurement. The chapter lists feedstock among the “exceptions” (p. 77) without flagging this risk.
2. Declared compliance needs independent verification. Covert production resumed after a legal phase-out and was caught only by long-term measurement networks, not by the reporting system. This extends pp. 81–82.
3. A legacy stock can become a resource that locks in continued use. Where alternatives are lacking for some uses, stocks slated for destruction become essential supply. Policy then shifts from elimination to conservation. This complicates p. 77 and p. 81.
4. Projected recovery dates tend to recede as hidden stocks and leaks are found. Early projections were optimistic by about one to two decades on some metrics. This extends p. 78.
5. A regime can adapt even when it starts late. The ozone regime detected its own failures, tightened repeatedly and extended its remit (to HFCs) by consensus. That tempers the chapter’s pessimism about consensus pace (p. 78, fn 6) without contradicting its account of the early years.
6. Classifying a decision as “precautionary” depends on the definition. Formal and political-economy readings of the same decision can both be defended (p. 80). A synthesis should say which definition it uses.

Method notes and limits#

• Not retrievable:
• DETR (1999), the source of Fig. 7.2 and of the claim that only 1995 secured recovery;
• the SPARC carbon tetrachloride report (site unreachable; cited via WMO 2022);
• the full texts of Benedick (1998, 2009) and Parson (2003);
• the Maxwell & Briscoe abstract and the April 1988 C&EN item on DuPont (publisher blocked access);
• a primary source for the USD 150 million China approval of March 1999;
• the full IISD Earth Negotiations Bulletin summary of MOP-37 (blocked access; used only through a search excerpt and the Ozone Secretariat’s own summary);
• the NOAA news release on the 2025 ozone hole (blocked access; the NASA version of the same release was used).

Where a verdict leans on these, it is flagged. - Resolved in the second pass: - Benedick’s “2005 testimony” (Senate EPW, 28 Sept 2005); - the original 1987 preamble wording; - China’s July 2007 CFC and halon plant closures; - WMO’s September 2025 Ozone and UV Bulletin. - Secondary sources used, and why: contemporaneous news reports (UPI 1988, People’s Daily 2000) were used for DuPont’s 1988 target and China’s 2000 halon plan, and a statistician’s blog (Hyndman 2023) for the historians’ account of the satellite data. No primary source for these points was reachable. - Not re-checked: the US production shares (46% and 28%), which Farman reports as a US claim; halon production dates for India and Korea; and the 791 kt CFC-12 bank figure for 1995. - Metric caution. Farman’s “equivalent chlorine” is a tropospheric index using 58 × bromine. WMO’s EESC is a stratospheric index with transport lags, and its formulation changed in 2018–2022. Return dates are not strictly comparable across these, and this check says so where it matters. - Scope. Hydrofluoroolefins, TFA, N2O and dichloromethane are included only because they are halocarbons or ozone-depleting substances within the chapter’s own subject. No technology outside the reports is discussed.

Sources#

Scientific assessments (WMO/UNEP, NOAA, TEAP, EEAP)

• WMO/UNEP, Scientific Assessment of Ozone Depletion: 2022, GAW Report No. 278, Executive Summary (pub. Jan 2023). https://csl.noaa.gov/assessments/ozone/2022/executivesummary/
• WMO/UNEP 2022, Chapter 1, “Update on Ozone-Depleting Substances (ODSs) and Other Gases of Interest to the Montreal Protocol”. https://csl.noaa.gov/assessments/ozone/2022/downloads/Chapter1_2022OzoneAssessment.pdf
• WMO/UNEP 2022, Chapter 7, “Scenarios and Information for Policymakers” (incl. Table 7-1, Table 7-6, Appendix Table 7A-1). https://csl.noaa.gov/assessments/ozone/2022/downloads/Chapter7_2022OzoneAssessment.pdf
• WMO/UNEP, Twenty Questions and Answers About the Ozone Layer: 2022 Update (Q14, Q16). https://csl.noaa.gov/assessments/ozone/2022/downloads/twentyquestions.pdf
• WMO/UNEP, Scientific Assessment of Ozone Depletion: 1994, Executive Summary. https://csl.noaa.gov/assessments/ozone/1994/executivesummary.html
• NOAA CSL, Ozone assessments index (“2026 assessment in preparation”), accessed 25 Sept 2026. https://csl.noaa.gov/assessments/ozone/
• NOAA CSL, Scientific Assessment of Ozone Depletion: 2026 page (core report due end of 2026), accessed 25 Sept 2026. https://csl.noaa.gov/assessments/ozone/2026/
• WMO, Ozone and UV Bulletin No. 3 (Sept 2025). https://ozone.unep.org/sites/default/files/2025-09/Ozone-and-UV-Bulletin_EN.pdf
• NASA Science, “NASA, NOAA Rank 2025 Ozone Hole as 5th Smallest Since 1992” (24 Nov 2025). https://science.nasa.gov/earth/nasa-noaa-rank-2025-ozone-hole-as-5th-smallest-since-1992/
• NOAA GML, Ozone-Depleting Gas Index, 2024 report (fall 2024). https://gml.noaa.gov/odgi/
• TEAP, 2022 Assessment Report of the Technology and Economic Assessment Panel (pub. Apr 2023). https://ozone.unep.org/system/files/documents/TEAP-Assessment-Report-2022-April23.pdf
• Neale R.E. et al. (EEAP), “The effects of exposure to solar radiation on human health”, Photochem. Photobiol. Sci., 1 Mar 2023. https://doi.org/10.1007/s43630-023-00375-8
• Neale P.J. et al. (EEAP), “Environmental consequences of interacting effects of changes in stratospheric ozone, UV radiation, and climate: Update 2024”, Photochem. Photobiol. Sci., 17 Mar 2025. https://doi.org/10.1007/s43630-025-00687-x
• SPARC, Report on the Mystery of Carbon Tetrachloride, SPARC Report No. 7, 2016 (cited via WMO 2022; not opened). https://www.wcrp-climate.org/WCRP-publications/2016/SPARC_Report7_2016.pdf

Treaty texts, decisions and implementation records

• Montreal Protocol, preamble (consolidated text), Ozone Secretariat. https://ozone.unep.org/treaties/montreal-protocol/articles/preamble
• Montreal Protocol, original 1987 text, Official Journal of the European Communities L 297/21 (31 Oct 1988). https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:21988A1031(02)
• Vienna Convention, preamble, Ozone Secretariat. https://ozone.unep.org/treaties/vienna-convention/articles/preamble
• Ozone Secretariat, Status of ratification (as of 10 Aug 2026). https://ozone.unep.org/all-ratifications
• Decision XIII/17, Compliance with the Montreal Protocol by the Russian Federation (2001). https://ozone.unep.org/treaties/montreal-protocol/meetings/thirteenth-meeting-parties/decisions/decision-xiii17-compliance-montreal-protocol-russian-federation
• Decision XIV/35, Compliance with the Montreal Protocol by the Russian Federation (2002). https://ozone.unep.org/treaties/montreal-protocol/meetings/fourteenth-meeting-parties/decisions/decision-xiv35-compliance-montreal-protocol-russian-federation
• Decision XIX/6, Adjustments … with regard to Annex C, Group I, substances (HCFCs) (Sept 2007). https://ozone.unep.org/treaties/montreal-protocol/meetings/nineteenth-meeting-parties/decisions/decision-xix6-adjustments-montreal-protocol-regard-annex-c-group-i-substances
• Decision XXX/3, Unexpected emissions of trichlorofluoromethane (CFC-11) (Nov 2018). https://ozone.unep.org/treaties/montreal-protocol/meetings/thirtieth-meeting-parties/decisions/decision-xxx3-unexpected-emissions-trichlorofluoromethane-cfc-11
• Decision XXX/7, Future availability of halons and their alternatives (Nov 2018). https://ozone.unep.org/treaties/montreal-protocol/meetings/thirtieth-meeting-parties/decisions/decision-xxx7-future-availability-halons-and-their-alternatives
• Decision XXXVI/2, Life-cycle refrigerant management (MOP-36, 2024). https://ozone.unep.org/treaties/montreal-protocol/meetings/thirty-sixth-meeting-parties/decisions/decision-xxxvi2-life-cycle-refrigerant-management
• Decision XXXVI/5, Feedstock uses of controlled substances (MOP-36, 2024). https://ozone.unep.org/treaties/montreal-protocol/meetings/thirty-sixth-meeting-parties/decisions/decision-xxxvi5-feedstock-uses-controlled-substances
• Ozone Secretariat, MOP-37 meeting summary (Nairobi, 3–7 Nov 2025), accessed 25 Sept 2026. https://ozone.unep.org/meeting-summary
• IISD Earth Negotiations Bulletin, MOP-37 summary report (Nov 2025; secondary; seen only as a search excerpt). https://enb.iisd.org/montreal-protocol-meeting-parties-ozone-mop37-summary
• World Bank, press release “China Closes Ozone Depleting Chemical Plants” (1 July 2007; archived copy of 17 Apr 2025). http://web.archive.org/web/20250417055132/https://www.worldbank.org/en/news/press-release/2007/07/01/china-closes-ozone-depleting-chemical-plants
• Multilateral Fund, homepage statistics (accessed 25 Sept 2026). https://www.multilateralfund.org/
• Multilateral Fund, “Our history” timeline (accessed 25 Sept 2026). https://www.multilateralfund.org/about/history
• World Bank, Implementation Completion and Results Report ICR00002241, India CFC Production Sector Gradual Phase-out Project (22 June 2012). http://documents.worldbank.org/curated/en/993301468041466573/text/ICR22410P069370C0disclosed070100120.txt
• World Bank, Montreal Protocol 2007 Business Plan (29 Jan 2007). http://documents.worldbank.org/curated/en/282211468161672032/text/418390WP0Montr10070321443B01PUBLIC1.txt
• European Commission, Protecting the ozone layer (accessed Sept 2026). https://climate.ec.europa.eu/areas-action/ozone-layer_en
• European Commission, ODS legislation / Regulation (EU) 2024/590 summary (accessed Sept 2026). https://climate.ec.europa.eu/eu-action/ozone-layer/ods-legislation_en
• Regulation (EU) 2024/590 on substances that deplete the ozone layer (EUR-Lex; could not be opened directly, content taken from the Commission summary). https://eur-lex.europa.eu/eli/reg/2024/590/oj/eng
• European Commission, F-gas legislation / Regulation (EU) 2024/573 (adopted 7 Feb 2024). https://climate.ec.europa.eu/areas-action/fluorinated-greenhouse-gases/f-gas-legislation_en
• US EPA, Health and Environmental Effects of Ozone Layer Depletion (citing May 2020 AHEF report). https://www.epa.gov/ozone-layer-protection/health-and-environmental-effects-ozone-layer-depletion
• US EPA, Reducing Hydrofluorocarbon (HFC) Use and Emissions (AIM Act page, accessed Sept 2026). https://www.epa.gov/climate-hfcs-reduction
• UN Treaty Collection, Kyoto Protocol status (accessed 25 Sept 2026). https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=XXVII-7-a&chapter=27&clang=_en
• UN Treaty Collection, Paris Agreement status, note 7 (US withdrawal effective 27 Jan 2026). https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=XXVII-7-d&chapter=27&clang=_en
• White House, Executive Order “Putting America First in International Environmental Agreements” (20 Jan 2025). https://www.whitehouse.gov/presidential-actions/2025/01/putting-america-first-in-international-environmental-agreements/
• White House, Memorandum “Withdrawing the United States from International Organizations, Conventions, and Treaties that Are Contrary to the Interests of the United States” (7 Jan 2026). https://www.whitehouse.gov/presidential-actions/2026/01/withdrawing-the-united-states-from-international-organizations-conventions-and-treaties-that-are-contrary-to-the-interests-of-the-united-states/

Peer-reviewed studies

• Slaper H. et al., “Estimates of ozone depletion and skin cancer incidence to examine the Vienna Convention achievements”, Nature 384:256, Nov 1996. https://doi.org/10.1038/384256a0
• Velders G.J.M. et al., “The importance of the Montreal Protocol in protecting climate”, PNAS, Mar 2007. https://doi.org/10.1073/pnas.0610328104
• Velders G.J.M. et al., “The large contribution of projected HFC emissions to future climate forcing”, PNAS, 22 June 2009. https://doi.org/10.1073/pnas.0902817106
• Ravishankara A.R., Daniel J.S., Portmann R.W., “Nitrous oxide (N2O): the dominant ozone-depleting substance emitted in the 21st century”, Science, 27 Aug 2009. https://doi.org/10.1126/science.1176985
• Rockström J. et al., “A safe operating space for humanity”, Nature, Sept 2009. https://doi.org/10.1038/461472a
• Newman P.A. & McKenzie R., “UV impacts avoided by the Montreal Protocol”, Photochem. Photobiol. Sci., Apr 2011. https://doi.org/10.1039/c0pp00387e
• van Dijk A. et al., “Skin cancer risks avoided by the Montreal Protocol — worldwide modeling integrating coupled climate-chemistry models with a risk model for UV”, Photochem. Photobiol., online 26 Sept 2012 (2013 issue). https://doi.org/10.1111/j.1751-1097.2012.01223.x
• Solomon S. et al., “Emergence of healing in the Antarctic ozone layer”, Science, 30 June 2016. https://doi.org/10.1126/science.aae0061
• Montzka S.A. et al., “An unexpected and persistent increase in global emissions of ozone-depleting CFC-11”, Nature, 16 May 2018. https://doi.org/10.1038/s41586-018-0106-2
• Rigby M. et al., “Increase in CFC-11 emissions from eastern China based on atmospheric observations”, Nature, 22 May 2019. https://doi.org/10.1038/s41586-019-1193-4
• Stanley K.M. et al., “Increase in global emissions of HFC-23 despite near-total expected reductions”, Nature Communications, 21 Jan 2020. https://doi.org/10.1038/s41467-019-13899-4
• Lickley M. et al., “Quantifying contributions of chlorofluorocarbon banks to emissions and impacts on the ozone layer and climate”, Nature Communications, 17 Mar 2020. https://doi.org/10.1038/s41467-020-15162-7
• Montzka S.A. et al., “A decline in global CFC-11 emissions during 2018−2019”, Nature, 10 Feb 2021. https://doi.org/10.1038/s41586-021-03260-5
• Park S. et al., “A decline in emissions of CFC-11 and related chemicals from eastern China”, Nature, 10 Feb 2021. https://doi.org/10.1038/s41586-021-03277-w
• Young P.J. et al., “The Montreal Protocol protects the terrestrial carbon sink”, Nature, 18 Aug 2021. https://doi.org/10.1038/s41586-021-03737-3
• Andersen S.O. et al., “Narrowing feedstock exemptions under the Montreal Protocol has multiple environmental benefits”, PNAS, Dec 2021. https://doi.org/10.1073/pnas.2022668118
• Western L.M. et al., “Global increase of ozone-depleting chlorofluorocarbons from 2010 to 2020”, Nature Geoscience, Apr 2023. https://doi.org/10.1038/s41561-023-01147-w
• Western L.M. et al., “A decrease in radiative forcing and equivalent effective chlorine from hydrochlorofluorocarbons”, Nature Climate Change, 11 June 2024. https://doi.org/10.1038/s41558-024-02038-7
• Wang P. et al., “Fingerprinting the recovery of Antarctic ozone”, Nature, 5 Mar 2025. https://doi.org/10.1038/s41586-025-08640-9
• Scheringer M., Arp H.P.H., Cousins I.T., “Boundaries, Limits, Global Threats — How Can the Impacts of Global Synthetic Pollutants Be Reduced?”, Environ. Sci. Technol., 2 Feb 2026. https://doi.org/10.1021/acs.est.5c13807
• Reimann S. et al., “Continuing industrial emissions are delaying the recovery of the stratospheric ozone layer”, Nature Communications, 16 Apr 2026. https://doi.org/10.1038/s41467-026-70533-w
• Holland R. et al., “TFA Generation and Deposition over Europe May Currently See a Greater Influence from HFO-1234yf than HFC-134a”, Environ. Sci. Technol. Lett., 15 June 2026. https://doi.org/10.1021/acs.estlett.6c00356

Participant testimony (primary)

• Benedick R.E., “The Case of the Montreal Protocol: Science Serving Public Policy”, testimony to the US Senate Committee on Environment and Public Works, 28 Sept 2005. https://www.epw.senate.gov/public/_cache/files/1/0/1094f8dd-dae7-45cd-b5a1-b9c018272318/BF9D594B66EBA773D15F23EC2FEC547786CB6ADB4C2DD1862C0C90B6D44D8B5A.092805benedick-testimony.pdf
• US Senate, S. Hrg. 109-1003, The Role of Science in Environmental Policy Making (hearing of 28 Sept 2005), GovInfo record. https://www.govinfo.gov/app/details/CHRG-109shrg38918

Contemporaneous news and secondary accounts (used only where no primary source was reachable)

• UPI, “The Ozone Dilemma: Du Pont leads movement to phase out compound that threatens ozone layer” (4 Sept 1988). https://www.upi.com/Archives/1988/09/04/The-Ozone-DilemmaNEWLNDu-Pont-leads-movement-to-phase-out-compound-that-threatens-ozone-layerNEWLNReplacing-chlorofluorocarbons-will-not-be-easy-inexpensive/8830045809758/
• C&EN, “CFC production: Du Pont seeks total phaseout”, vol. 66 no. 14 (Apr 1988; not opened). https://pubs.acs.org/doi/abs/10.1021/cen-v066n014.p004
• People’s Daily (English), “China Makes Progress in Reducing Use of Halon” (23 Nov 2000). https://en.people.cn/english/200011/23/print20001123_55903.html
• Hyndman R.J., “How NASA didn’t discover the hole in the ozone layer” (blog, 29 Sept 2023), summarising Christie M., The Ozone Layer: A Philosophy of Science Perspective (CUP, 2001) and Pukelsheim F., “Robustness of statistical gossip and the Antarctic ozone hole”, IMS Bulletin 19(4), 1990. https://robjhyndman.com/hyndsight/ozone-hole-anomaly.html

Histories and participant accounts (bibliographic records; full texts not re-read in this pass)

• Benedick R.E., Ozone Diplomacy: New Directions in Safeguarding the Planet, enlarged ed., Harvard University Press, 1998 (1st ed. 1991). https://doi.org/10.4159/9780674020757
• Benedick R.E., “Science Inspiring Diplomacy: The Improbable Montreal Protocol”, in Zerefos C. et al. (eds), Twenty Years of Ozone Decline, Springer, 2009. https://doi.org/10.1007/978-90-481-2469-5_4
• Parson E.A., Protecting the Ozone Layer: Science and Strategy, Oxford University Press, 2003. https://doi.org/10.1093/0195155491.001.0001
• Maxwell J. & Briscoe F., “There’s money in the air: the CFC ban and DuPont’s regulatory strategy”, Business Strategy and the Environment 6(5):276–286, Nov 1997. https://doi.org/10.1002/(SICI)1099-0836(199711)6:5<276::AID-BSE123>3.0.CO;2-A

Later EEA treatment

• EEA, Late lessons from early warnings: science, precaution, innovation, EEA Report No 1/2013 (22 Jan 2013): Annex 2, p. 708 (case summary and Table A2.7); Annex 3, p. 728 (“Ozone layer update”, a summary of WMO Twenty Questions: 2006); Ch. 23, Box 23.3, pp. 575–576; Ch. 26, fn 4, p. 631; Ch. 3, p. 50. Publication page: https://www.eea.europa.eu/en/analysis/publications/late-lessons-2 ; full report: https://doi.org/10.2800/73322