Late Lessons, Jensen Huang and AI
Markdown version

LL1-10 hindsight check: Sulphur dioxide (Semb), Late lessons from early warnings (EEA, 2001), Ch. 10, pp. 101–109#

Checked 25 September 2026. The check covers what happened between 2001 and September 2026. Where they help weigh a claim, it also uses a few older primary sources that the chapter did not use, such as Hansard and the emissions table behind the chapter’s own Figure 10.1. Page numbers refer to the 2001 report.

Overview#

The chapter tells a historical story and makes a handful of live claims and forecasts. The core of the story has held up well: tall stacks moved sulphur pollution rather than reducing it; Scandinavian acidification was traced to other countries’ emissions; the regime built under the transboundary air pollution convention (CLRTAP) worked. Several of its numbers and comparisons do not survive checking: the tenfold rise, the London toll and the limits comparison. One uncited claim, on maintenance savings, could not be traced. One forecast, on forest vitality, went the wrong way.

Vindicated or strengthened - Dispersion instead of reduction. Mylona’s (1996) inventory is the source of Figure 10.1. It shows European emissions rising from 23.4 Mt (1950) to a peak of 57.4 Mt (1975). Later retrospectives treat the tall-stack policy as the textbook case of dilution that failed (Fowler et al., 2020; Grennfelt et al., 2020). - Source–receptor science and critical loads. The “blame matrices” became routine EMEP output and were extended to nitrogen and ozone. In the EU-28, the ecosystem area where critical loads for acidity were exceeded fell from 43% (1980) to 7% (2010) (EEA). EMEP’s 2026 status report puts it at 3.5% of Europe’s mapped ecosystem area in 2024. That is well beyond the “60 % gap closure” the chapter says was achievable (p. 106). - Emissions and lakes. SO2 emissions fell far further than the chapter could report: EU-wide down 84% (2005–2023), UK down 98% since 1990, Germany down 96% since 1990. Chemical recovery of lakes is now overwhelmingly documented. Biological recovery lags by decades, a lag the chapter’s “slow, but steady” (p. 106) understates. - The German advisers’ caution (pp. 105, 107). SO2 cuts did not cure the widespread “new forest damage”. German crown thinning was worse in 2025 (significant thinning on 35% of forest area) than in 1984 (23%, West Germany), now driven by drought and heat. SO2 was a proven cause only in the heavily polluted border mountains of the “Black Triangle”, and those forests recovered as SO2 fell. - The shift in emphasis to nitrogen, ammonia, ozone and eutrophication (p. 106). This forecast was correct. In 2024 the critical load for eutrophication was still exceeded on 60.6% of Europe’s ecosystem area. EU ammonia emissions fell only 17% between 2005 and 2023.

Wrong, overstated or unverifiable - The tenfold rise in GDR, Czechoslovak and Polish emissions, 1960–1985 (p. 104). This is not supported by the chapter’s own source. Mylona’s table shows their combined emissions roughly doubling, from 6.6 to 12.8 Mt (my sums). The second half of the sentence is correct: in 1985 these three countries emitted more than all of northwest Europe combined (12.8 vs about 9.7 Mt). - “Forest vitality is improving” (p. 106). This did not hold as a trend. ICP Forests reports that mean defoliation across Europe has risen since 1990. - Retrospective studies showing abatement costs “more than been recovered” through lower maintenance costs (p. 106). I could not find or verify these studies. The nearest candidate I found is a prospective European estimate (Cowell and ApSimon, 1996). It says building savings “may compensate for a considerable proportion of abatement costs”, which is weaker than “more than recovered”. Corrosion of materials did fall with SO2 (ICP Materials). But every later cost–benefit assessment I found (US and European, including the EEA’s own 2013 volume, Ch. 23, pp. 572–573) concludes that the benefits of SO2 control come overwhelmingly from fewer deaths caused by fine particles (sulphate is a major component), a channel the chapter never mentions. - London 1952 (p. 101). “More than 2 000” deaths in one week badly understates the toll. It matches the total deaths registered in the County of London in the week ending 13 December 1952 (2,484; Scott, 1953), not the event’s toll. The long-standing official-era figure is about 4,000 (the Met Office’s “known” deaths). A 2001 re-analysis puts excess deaths from December 1952 to February 1953 at about 12,000. The EEA’s own 2013 volume uses “as many as 13 000” (Ch. 23, p. 572), while its Annex 2 repeats “more than 2 000” (p. 711). - Air-quality limits “not very much lower” than the smog thresholds (p. 106). The comparison sets a 1-hour limit against thresholds that WHO’s criteria document gives as 24-hour means (WHO, 1979). Like for like, the 1999 daily limit was a quarter of the smog-derived threshold, and the 2030 limit is a tenth. EU compliance is now essentially complete: below 0.1% of the EU urban population has been exposed above the daily limit every year since 2010 (EEA, 2026). - Minor errors. Several facts in the chapter are slightly wrong (see “Minor factual checks”). EMEP began in 1977, not 1976. The UK and Poland were not the only countries that did not sign the 1985 protocol.

What the chapter could not see (post-2001 developments that change the frame) - Health took over as the main policy driver around 2000. When the Gothenburg Protocol was revised in 2012, fine particles were included in an international protocol for the first time (Grennfelt et al., 2020). - Successful SO2 control had an unintended climate cost. Sulphate aerosols had been cooling Europe. One modelling study estimates that European air-quality measures between 1970 and 2010 warmed Europe by about 0.45 °C while preventing about 80,000 premature deaths a year (Turnock et al., 2016). - Formal commitments did not deliver cuts in the East. The GDR and Czechoslovakia signed the 1985 30% protocol while their emissions were still rising (Mylona, 1996; UN treaty record). Large reductions came only with economic collapse after 1990 (Vestreng et al., 2007; Umweltbundesamt).

Weight for the lens. The chapter’s strongest lessons are about mechanisms: - dispersion that relocates harm; - jointly produced source–receptor evidence; - measurable intermediate thresholds (critical loads); - long-term monitoring; - structural shifts doing much of the work.

The later record supports all five, several strongly. The quantitative asides (the tenfold rise, the London toll, maintenance savings, the limits-versus-thresholds comparison) should not be cited without correction. The case’s most useful new lesson from hindsight is not in the chapter at all. The largest benefits of controlling this pollutant came through a harm (fine-particle mortality) that was not the reason for acting, and so did the largest unintended cost (unmasked warming).

2013 update. Annex 3 of Late lessons II (EEA Report 1/2013, “An update of some case studies”, pp. 717 ff.) does not update the SO2 case. Its updates cover fisheries, benzene, asbestos, PCBs, halocarbons, DES, antimicrobials, MTBE and hormones. The 2013 volume touches the case in four other places (all retrieved from the EEA website): - Summary (p. 24): acid rain is among the “more established issues” of the 2001 volume whose lessons were “valuable and robust”. - Annex 2 (pp. 702, 711) restates the 2001 chapter without revision. Table A2.10 repeats “more than 2 000” (1952), 800 (1962) and the garbled 1988 directive line. Table A2.1 (after Gee, 2009) dates the early warnings to 1952 (lungs) and 1968 (lakes), effective action to 1979–2001, and puts the “years of substantial inaction” at 25–55. It projects “c. 90 % reduction on 1975 levels by 2010”. - Chapter 23 (Andersen and Clubb, Box 23.2, pp. 572–573) is in effect the real update. It moves the case onto health costs: the 1952 smog caused “as many as 13 000” excess deaths; the health costs of inaction “should indeed have been counted in billions”. It cites a European Commission cost–benefit analysis (AEA, 2005) that put the cost of air pollution in the EU-25 at EUR 276–427 billion a year (3–5% of GDP). It also cites a damage cost of EUR 5–9 per kg SO2 against marginal abatement costs “from below EUR 1 per kg”. - Chapter 2 (Hansen and Tickner, p. 20 and Table 2.3) looks at acid rain as one of 88 alleged “false positives” of regulation named by critics (Bast et al., 1994; Wildavsky, 1995). It classifies acid rain as a “real risk”, on the strength of the US NAPAP assessment (1996).

Claim 1. The December 1952 London smog killed “more than 2 000 people” in one week; a 1962 smog killed “about 800” more (p. 101; Table 10.1, p. 108)#

Original claim. - “Records from London’s hospitals” showed that “more than 2 000 people died” from the exposure “during one week” (p. 101). - A second smog in 1962 caused “about 800 additional deaths” (p. 101). - Table 10.1 repeats both figures (p. 108). - The chapter cites Brimblecombe (1987) and Ashby and Anderson (1981) for the context.

What happened since. - The long-standing figure. The Met Office’s current case study says “about 4,000 people were known to have died as a result of the fog, but it could be many more”. For 1962 it gives 750 deaths. - Where “more than 2 000” probably comes from. The London County Council’s medical officer of health reported to his committee on 27 January 1953 (Scott, 1953, Public Health Reports). He gave these deaths registered weekly in the administrative County of London: - 945 in the week ending 6 December; - 2,484 in the week ending 13 December; - 1,523 the following week; - 1,372 and 1,216 in the first two weeks of January 1953.

A normal winter week reached “as many as 1,100”. So the chapter’s figure matches all deaths registered in the inner county in the peak week, not deaths caused by the smog (the excess over the previous week was about 1,500). - For the wider Greater London area, the widely reported Registrar General’s figure for the same week is 4,703 deaths, against 1,852 in the same week of 1951. That is an excess of roughly 2,850 in one week. I could not trace this to a primary document, and I could not read Logan (1953, Lancet), the standard contemporary analysis. - Either way, the chapter’s number is a one-week count and not the event’s toll, and the chapter does not say so. - The mortality count came from death registrations, not “records from London’s hospitals”. That is a minor slip. - The 2001 re-analysis. Bell and Davis (2001, Environmental Health Perspectives) used insurance claims, hospital admissions, pneumonia reports, mortality and influenza data. - Mortality from December 1952 to February 1953 was 50–300% above the previous year. - They estimated about 12,000 excess deaths from acute and persisting effects. - They rejected the official view that influenza explained the raised deaths of early 1953. - The influenza question. Bell, Davis and Fletcher (2004) tested the influenza explanation directly. Influenza could account for only a fraction of the excess deaths. To explain most of them would require an epidemic with about twice the case-fatality rate and four times the incidence seen in general practice that winter. - Uptake. The 12,000 figure has since been adopted in authoritative reviews, for example Fowler et al. (2020, Philosophical Transactions A): “approximately 12 000 people”. - The EEA’s own later usage is inconsistent. In 2013 the EEA’s cost-of-inaction chapter gives “an excess mortality of as many as 13 000 deaths”, citing Bell, Davis and Fletcher (Andersen and Clubb, 2013, Box 23.2, p. 572). That paper’s abstract gives no total, so the 13,000 is slightly above the 12,000 its authors published in 2001. In the same volume, Annex 2 (Table A2.10, p. 711) repeats the 2001 chapter’s “more than 2 000”. - Independence. I found no independent re-derivation that contradicts it. The figure comes from one research group’s method (comparing 1952–53 with the previous winter) and should be cited as an estimate.

Verdict: strengthened. The harm was far larger than the chapter says. “More than 2,000” is accurate only as a one-week count: it matches all deaths registered in the County of London in the peak week (2,484; Scott, 1953). It is not a count of deaths caused by the smog, nor the event’s toll. The 1962 figure (about 800 against the Met Office’s 750) holds.

Weight. Do not quote “more than 2,000” as the toll. Use “about 4,000 in the immediate episode (official-era estimate), and up to about 12,000 excess deaths over three months (Bell and Davis, 2001)”. The correction strengthens the chapter’s framing that the smog was a public-health catastrophe. It also sharpens a lesson the chapter does not draw: the official account of the day attributed part of the toll to another cause (influenza). The full size of the harm was established only half a century later.

Claim 2. Tall stacks improved ground-level air while European SO2 emissions rose from about 23 Mt (1950) to a peak of about 57 Mt (mid-1970s); the industry’s confidence in dilution was wrong (pp. 101–103, Fig. 10.1)#

Original claim. - Under “best practicable means”, the main response to SO2 was tall chimneys “in proportion to the emitted amounts” (p. 101). - Surface air quality improved “in spite of the increased emissions”. The electricity industry was “confident that the emissions could be diluted and dispersed to levels that were not harmful”. “This was not so” (p. 102). - Figure 10.1 (p. 103; Semb, based on Mylona, 1996) shows European emissions from 1880 to 1990.

What happened since. - The chapter’s figures match its source. I checked Figure 10.1 against Table 1 of Mylona (1996, Tellus B 48B: 670–671), retrieved from the journal archive. Totals for Europe (including the European part of the USSR within the EMEP domain): - 23.4 Mt (1950); - 38.8 Mt (1960); - 55.5 Mt (1970); - 57.4 Mt (1975, the peak); - 55.1 Mt (1980); - 50.4 Mt (1985); - 48.2 Mt (1990).

Mylona’s abstract gives the peak as “approximately 55 million tonnes” in the 1970s and uncertainties of ±30–45% for national estimates. - Later inventories agree on the shape. The magnitude differs with the domain chosen: - Vestreng et al. (2007, Atmospheric Chemistry and Physics), using emissions reported to EMEP, give about 55 Tg SO2 in 1980, falling to 15 Tg in 2004. - Fowler et al. (2020) put Europe’s peak at 32 Mt of sulphur (about 64 Mt SO2) around 1970–1980. - The dilution critique is now standard history. - Grennfelt et al. (2020, Ambio), a review by acid-rain scientists, report that when Sweden first raised acid rain at the OECD in 1967, the committee’s view was that SO2 “was a local problem, which easily could be solved by tall stacks”. - Fowler et al. (2020) describe new UK power stations built from the 1950s with stacks of about 200 m to promote dispersion. Urban air improved “while country-scale emissions of SO2 were close to their maximum”. - They also report that the UK knew of domestic rural damage (failed pine plantations in the Pennines, lichen loss, crop effects). It was judged “insufficient” for further legislation. - In the UK the first flue-gas desulphurisation at a large power station was at Drax, from 1988 (Fowler et al., 2020). - UK emissions peaked early. In Mylona’s table they peaked at 6.5 Mt in 1965. That is consistent with the chapter’s point that urban clean-up and national emissions moved independently.

Verdict: held up. The figures are accurately drawn from Mylona. The central point, that dispersion improved the local indicator while total emissions grew and moved the harm downwind, is now the standard account.

Weight. This is the chapter’s most robust lesson (digest insight 1: a fix that disperses or relocates harm improves the visible indicator while total harm grows). Grade it strong. When citing the numbers, name the domain: the European totals include the European USSR.

Original claim. - The OECD report (1977) established exports and imports of sulphur between countries. It made clear that acidification in Scandinavia “could be quantitatively related to emissions in several European countries” (p. 103). - The Norwegian SNSF programme (1972–80) documented “widespread acidification of rivers and lakes” and “declining stocks of both trout and salmon” by 1976 (p. 103). - Forest effects were “much more difficult to demonstrate” (p. 103).

What happened since. - Source–receptor accounting became the regime’s core tool. - Grennfelt et al. (2020) and Hov and Grennfelt (2026, Ambio) describe the OECD report’s “blame matrices” as the first of the convention’s “bridging concepts”. Tabulated matrices became standard annual EMEP output. - They were later extended to oxidised nitrogen, ammonia and ozone (Hov and Grennfelt, 2026). - Both reviews are by participants (Grennfelt and Hov worked in the programmes) and should be read as insiders’ accounts. - Fish damage was confirmed and its mechanism clarified. Rosseland (2021, Ambio) summarises: - aluminium was identified around 1977 as the key toxic agent in acid water; - Atlantic salmon smolts turned out to be the most sensitive life stage; - large-scale liming programmes have kept many rivers habitable; in 2017, 15% of wild salmon caught by anglers in Norway came from limed rivers. - The forest finding stood. Grennfelt et al. (2020) and Fowler et al. (2020) both record that the SNSF studies found no significant evidence of acid-rain effects on Norwegian forests. In January 1985 a UK minister made the same distinction in the Commons: the UK had “a case … to answer” on Scandinavian freshwaters, but forest damage was a different question (HC Deb 11 January 1985, cc1021–1023). - An early-warning pre-history the chapter omits. Grennfelt et al. (2020) note that Robert Angus Smith described “acid rain” in 1872. Eville Gorham’s work (1955–63) laid much of the foundation. Both were largely ignored, meeting what Gorham called a “thundering silence”.

Verdict: held up. Every element has been confirmed and built on. The one qualification comes from the pre-history. The 1967–68 warning was not the first; earlier warnings were available and ignored.

Weight. Strong support for digest insight 7 (jointly produced source–receptor knowledge turns contested blame into allocatable obligations) and insight 6 (long-term monitoring detects slow, diffuse change). The Smith and Gorham history adds a point the chapter underplays. The problem was less that knowledge was missing than that it went unused, which qualifies Semb’s “limited understanding” (see Claim 6).

Claim 4. The 1981 OECD study found fish and timber losses economically small but materials damage comparable to abatement costs; uncited “retrospective studies” show abatement costs “more than been recovered” through lower maintenance costs (pp. 104, 106)#

Original claim. - OECD (1981) confirmed that lost fish were worth little and timber losses were “not particularly significant”. Damage to buildings and materials entailed costs “comparable” to those of flue-gas cleaning and fuel desulphurisation (p. 104). - Later: “a number of retrospective studies” showed that abatement costs “have more than been recovered” through lower maintenance of buildings and steel in cities (p. 106). No study is cited. - The Norwegian reply to the 1977 Nature editorial argued that a complete cost–benefit analysis would include damage in the emitting countries themselves (pp. 103–104).

What happened since. - Materials damage did fall with SO2. The convention’s materials programme (ICP Materials) reports 1987–2014 data from 19 European sites (Tidblad et al., 2017, Materials). - Corrosion and pollution “decreased significantly”, with a sharp fall until about 1997, then slower change. - Average urban SO2 at the sites fell below 10 µg/m³ by 1997 and below 5 µg/m³ by 2005. - The physical link the chapter relies on is confirmed. - The “more than recovered” claim remains unverified. I found no post-2001 study showing that SO2 abatement paid for itself through maintenance savings alone. The nearest pre-2001 work is a pair of 1996 papers by Cowell and ApSimon. - Cowell and ApSimon (1996, Atmospheric Environment) adapted a Nordic method (case studies of Stockholm, Prague and Sarpsborg) to estimate building savings under the 1994 Oslo protocol. Their abstract concludes that savings “are potentially very large, particularly in Eastern Europe, and may compensate for a considerable proportion of abatement costs”. - ApSimon and Cowell (1996, Energy Policy) describe the method. Their abstract reports no result. - These are prospective model estimates, not retrospective studies, and they claim “a considerable proportion”, not full recovery. If this literature is Semb’s basis, the chapter overstates it. - The papers’ own policy point is still worth noting. Because building damage is largely local, counting it shifts the case for abatement “from transboundary to domestic” benefits. That supports the Norwegian “Europe versus itself” argument (pp. 103–104). - Later cost–benefit work changes the ground of the argument. In every assessment found, health benefits from lower fine-particle (PM2.5) exposure dominate. Sulphate formed from SO2 is a major part of PM2.5. - US Acid Rain Program (Chestnut and Mills, 2005, Journal of Environmental Management): - benefits over $100 billion a year by 2010, mostly PM2.5 health benefits; - annualised costs about $3 billion, less than half the 1990 estimate; - ecosystem problems proved “more challenging to resolve than originally thought”. - US EPA, 1990 Clean Air Act Amendments (second prospective study, 2011): central benefits exceed costs by more than 30 to 1; about 85% of benefits come from reduced mortality linked to particulate matter. - Europe, retrospective (Turnock et al., 2016, Environmental Research Letters). Legislative and technology measures from 1970 to 2010: - cut 2010 SO2 emissions by 53%; - cut sulphate concentrations by 44%; - prevented about 80,000 premature deaths a year in the EU (95% range 37,000–116,000), valued at about $232 billion a year (1.4% of EU GDP). - Europe, industrial pollution (EEA, 2024 update of its industrial-emissions cost series): external costs of €268–428 billion a year over 2012–2021. That is still large, but down about 35% over the decade. - The EEA’s own 2013 reframing. Late lessons II revisits the case under “costs of inaction” (Andersen and Clubb, 2013, Box 23.2, pp. 572–573). - It quotes the same 1977 “million dollar problem with a billion dollar solution” line the 2001 chapter cites. It answers that once mortality is counted, “the health costs of inaction on air pollution damages should indeed have been counted in billions too”. - It cites the European Commission’s 2005 cost–benefit analysis for the Thematic Strategy on Air Pollution (AEA, 2005): EUR 276–427 billion a year for the EU-25 (3–5% of GDP), and 15–22% of GDP in Poland and other new member states. - It cites a damage cost of EUR 5–9 per kg SO2 against marginal abatement costs “from below EUR 1 per kg” (Rive, 2010). - In short, the EEA’s later volume makes the health-benefit argument that Semb’s chapter lacks. It does not repeat the maintenance-savings claim. - Fish and forests. The 1981 judgement that fish losses were small in money terms was never overturned. But later assessments treat ecosystem damage as poorly captured by market values (Fowler et al., 2020). Critical loads (Claim 8) were adopted partly because the damage could not be credibly priced (p. 107).

Verdict: partly held up. - The direction is strengthened: once all damages are counted, abatement pays many times over. The Norwegian “Europe versus itself” argument (pp. 103–104) proved right. - The mechanism the chapter offers (maintenance savings) is unverified and, by later accounting, minor. - The benefit that actually carried the economic case (avoided deaths from fine particles) is absent from the chapter.

Weight. Digest insight 3 (the boundaries of an appraisal can decide its conclusion) is strongly reinforced, but in a way Semb did not foresee: the decisive boundary was health, not materials. Do not cite the “more than recovered” sentence as evidence. It is uncited. The nearest literature I found (Cowell and ApSimon, 1996) is prospective and claims only “a considerable proportion”.

Claim 5. SO2 emissions in the GDR, Czechoslovakia and Poland “increased by a factor of 10 from 1960 to 1985”, when these three emitted more than all other northwest European countries together (p. 104)#

Original claim. “The SO2 emissions in these countries increased by a factor of 10 from 1960 to 1985, when these three countries alone emitted more SO2 than all other countries in northwest Europe together” (p. 104). No source is given. The chapter’s Figure 10.1 draws on Mylona (1996).

What happened since. The check uses Mylona (1996), Table 1 (the chapter’s own source, kilotonnes SO2 a year; the sums are mine):

Country 1950 1960 1970 1980 1985 1960–1985 ratio
GDR 1,970 3,584 4,362 4,446 5,294 1.5×
Czechoslovakia 662 1,164 2,561 3,094 3,107 2.7×
Poland 1,043 1,805 2,730 4,114 4,373 2.4×
Three combined 3,675 6,553 9,653 11,654 12,774 1.9×

Verdict: partly held up. The first half (tenfold) is contradicted by the chapter’s own source. The second half (more than northwest Europe combined) is confirmed.

Weight. Drop “tenfold”. The substantive point survives: by the mid-1980s the largest sources were in states that could not or would not pay for abatement (p. 107). The accurate figures also make the chapter’s point about dependence on one fuel more precise. The eastern rise was steady growth on low-grade, high-sulphur lignite, not an explosion.

Claim 6. Action came only on “proof beyond reasonable doubt”; the precautionary approach was “altogether absent” early on because of decision-makers’ “limited understanding”; the UK moved only when acid rain’s “costs closer to home” were appreciated (pp. 106–107; the editors date UK acknowledgement to 1985, p. 180)#

Original claim. - “Generally action has only been taken on the basis of proof beyond reasonable doubt” (p. 106). - Precaution was “altogether absent, mainly because decision-makers had limited understanding of the problems” (p. 107). - The UK “arguably” changed only when acid rain “had costs closer to home” (p. 107). The chapter gives no date. - The CEGB chairman was “impelled” to tell the Prime Minister that Norway’s accusations were “scientifically well founded”. This comes after a “somewhat superfluous” Royal Society–Scandinavian research programme concluded that acid rain killed fish (Mason, 1992) (p. 105).

What happened since. Primary parliamentary records (Hansard) show the UK sequence clearly: - 11 December 1984. The government announced an expert group to assess “the extent of acidification of freshwaters in Scotland, England and Wales and the causes” (HC Deb 18 December 1984, c79W, referring to the announcement of 11 December). - 11 January 1985. In the Commons debate on the Environment Committee’s acid rain report, Environment Minister William Waldegrave: - accepted that there was “a case which the United Kingdom must answer” on freshwater acidification in southern Scandinavia; - said a relationship between emissions and deposition was “beginning to” be established; - argued that “it is right to wait to see what we are likely to get for our money”, because the environmental gain from reducing deposition had not been modelled (HC Deb 11 January 1985, cc1021–1022).

This supports the editors’ 1985 date for acknowledgement (p. 180). It also illustrates the demand for a proven dose–response before action that the chapter describes. - 11 September 1986. Environment Secretary Nicholas Ridley announced “the Government’s decision in principle” to fit flue-gas desulphurisation to three major existing power stations (HC Deb 19 November 1986, c552). - On 29 October 1986, a Labour peer referred in the Lords to the government’s now admitting “the link between sulphur emissions and acid deposition”. The minister did not dispute it (HL Deb 29 October 1986, c709). - Ministers continued to stress the UK’s falling emissions and the much higher per-head emissions of East Germany (HC Deb 19 November 1986, c553; HL Deb 29 October 1986, cc710–711). - They also resisted joining the “30% club”. - 24 November 1988. The EC Large Combustion Plants Directive (88/609/EEC) was adopted. The chapter’s Table 10.1 (p. 108) garbles this date. - 1994. The UK signed the Oslo protocol on 14 June 1994 and ratified it on 17 December 1996 (UN Treaty Collection). - The research programme. The Royal Society–Scandinavian Surface Waters Acidification Programme (SWAP) reported in 1990 (Mason, ed., Cambridge University Press). The UK’s decision in principle on desulphurisation (September 1986) therefore came before SWAP’s conclusions, not after. The chapter’s narrative leaves the order unclear. - “Costs closer to home.” - The domestic freshwater review (1984) and the Galloway evidence cited by ministers in 1986 fit the chapter’s reading. - So does Fowler et al.’s (2020) account of domestic damage that was known but not acted on. - That domestic evidence mattered is plausible, but I found no source showing it was decisive. - “Limited understanding” is only partly right. - Understanding was limited in the sense that forest causation was never settled (Claim 7). - But the long-range transport of sulphur had been quantified by 1977, and earlier warnings (Smith, 1872; Gorham, 1955–63) had been ignored (Grennfelt et al., 2020). - Ministers in 1985 accepted the freshwater case yet deferred action on cost grounds. - The record fits Semb’s other explanation better: positions followed perceived costs and benefits (p. 107). - “Proof beyond reasonable doubt” was not universal. West Germany acted in 1983 without an agreed cause for forest damage. Semb himself says the decision “had already been taken” (p. 107). So the “proof” rule describes the UK (and the early Eastern bloc), not Europe as a whole.

Verdict: partly held up. - The UK timeline supports the general claim of evidence demands and cost-driven positions, and the editors’ 1985 date for acknowledgement. - “Limited understanding” is weakened as the main explanation. The UK acknowledged the case in January 1985 but deferred action until September 1986 on grounds of value for money. - “Proof beyond reasonable doubt” does not fit the German case, as the chapter itself shows.

Weight. Use this case for digest insight 2 (positions on evidence follow who pays and who suffers), now better documented than the chapter shows: grade it moderate to strong. Treat Semb’s “limited understanding” as one participant’s charitable reading. The primary record shows acceptance of the evidence combined with a demand to know “what we are likely to get for our money” before acting.

Claim 7. The German Council of Environmental Advisers (1983) warned that SO2 cuts might not cure the “new forest damage”, and that justifying action on that basis risked the credibility of environmental policy (pp. 105, 107)#

Original claim. - The Council backed desulphurisation, but “was not convinced by the scientific explanation of the new forest damage”. - It cautioned that SO2 reductions “might not improve this particular situation”. Action justified on this basis alone “might jeopardise the credibility of the environmental policy” (p. 105). - It advocated proportionate reduction of “the total spectrum” of pollutants (p. 105). - Semb: the argument was “for the record”; politically the decision “had already been taken” (p. 107).

What happened since. - The widespread decline did not occur as forecast, and SO2 was not its general cause. - Kandler and Innes (1995, Environmental Pollution), a pre-2001 source the chapter did not cite, found that surveys “failed to confirm” forest death or decline over large areas of Central Europe. Tree growth rates were higher than ever recorded. They recommended dropping the terms “Waldsterben” and “neuartige Waldschäden”. - Pretzsch et al. (2014, Nature Communications) found Central European spruce and beech growing markedly faster than in 1960 (tree growth up 32–77%), mainly because of warming and longer growing seasons. - Fowler et al. (2020) record that there is still “no consensus to date” on how much of the 1980s damage each pollutant caused. - Grennfelt et al. (2020), acid-rain insiders, call the forest alarms “often exaggerated”. - Crown condition got worse as SO2 fell. - German national survey (BMLEH, Waldzustandserhebung 2025; West Germany only until 1989): - mean crown thinning was 18.9% in 1984, 18.8% in 2001 and 25.2% in 2025; - the share of forest area with significant thinning rose from 23% (1984) to 35% (2025); - the ministry attributes recent damage to the 2018–2020 droughts and heat, with insects and fungi. - Over the same period German SO2 emissions fell from 5.5 Mt (1990) to 0.2 Mt (2024) (Umweltbundesamt). - Across Europe, ICP Forests (40th anniversary report, 2025) reports that mean defoliation has increased since 1990. From 2010 to 2023, crown condition declined on 33.7% of plots and improved on 10.7%. It also notes that the link between 1980s damage and air pollution “was largely debated”. - Where SO2 was the cause, cutting it worked. - In the Czech “Black Triangle” and Ore Mountains, spruce tree-ring growth fell in the 1970s and recovered in the 1990s, “tracking SO2 concentrations closely”. Liming made no difference (Hruška et al., 2023, PLOS ONE; Kolář et al., 2015). - In the Ore Mountains, young spruce health is now at the Czech national average, but soils remain acidified and nutrient-poor (Novotný et al., 2024). - This matches the Council’s distinction between proven local SO2 damage and the unexplained general symptoms. - Credibility. I found no study that measures whether the episode damaged the credibility of environmental policy. - Historians treat “Waldsterben” as a historical phenomenon in its own right. Metzger, Bemmann and Schäfer (2007) note the counter-reading of it as “a purely media phenomenon” (Holzberger). See also Metzger (2015) and von Detten (ed., 2013). - The episode is widely cited as an exaggerated alarm, which is the risk the Council foresaw. - Critics of precautionary regulation did use it this way. Acid rain appears on lists of alleged regulatory “false positives”. As Hansen and Tickner summarise them, Bast et al. (1994) and Wildavsky (1995) argued that it posed little or no threat to forests, crops, health or lakes in the United States. I did not read those two sources myself. The EEA’s 2013 volume examined those lists and classified acid rain as a “real risk”, not a false alarm (Hansen and Tickner, 2013, p. 20 and Table 2.3). The basis was the US NAPAP assessment (1996): most forests were not then known to be harmed, but sensitive high-elevation spruce–fir forests in the eastern United States were. That is the same split between general and local damage that the German Council drew. - Yet the policies it helped trigger were not reversed. They were extended, and they delivered large health and freshwater benefits (Claims 4 and 8).

Verdict: strengthened on the science: the Council’s caution was borne out. Unclear on credibility. Overclaiming left a lasting “false alarm” reputation, and critics of precaution used it. But I found no evidence that it weakened support for European air policy.

Weight. Digest insight 5 (action can come before causal consensus; overclaiming a cause risks credibility; broad, proportionate justification is safer) moves from moderate to moderate–strong. The case now shows both halves: - the broad justification (reduce the whole spectrum of pollutants in proportion) was vindicated by benefits no one was counting in 1983; - the narrow justification (SO2 is killing the forests) was not.

The monitoring network set up because of the forest scare (ICP Forests, 1985) later became Europe’s main tool for detecting drought-driven dieback (insight 6).

Claim 8. Critical loads shifted the burden of proof and gave “a rational basis” for the 1994 sulphur protocol, but only “60 % gap closure” was achievable at the worst-affected sites (pp. 106–107)#

Original claim. - Critical loads, defined as “the highest deposition … that does not change the ecosystem in an unacceptable way”, were mapped across Europe and provided “a rational basis” for the 1994 protocol (p. 106). - Only “60 % gap closure” was achievable where exceedances were highest (p. 106). - The approach shifts the burden of proof from monetising future damage to the “relatively simple task” of setting acceptable deposition (p. 107). It is “not exactly” precaution but “a rational way of dealing with uncertainties” (p. 107).

What happened since. - Exceedances fell far below the 1994 ambition. - EEA indicator CSI 005 (assessment of 2 October 2018): the share of EU-28 ecosystem area exceeding acidity critical loads fell from 43% (1980) to 7% (2010), with 4% projected for 2020. - EMEP Status Report 1/2026 (24 August 2026): acidity exceedance on 3.5% of Europe’s mapped ecosystem area in 2024, with “hot spots” no longer identifiable. The EMEP domain (about 4 million ecosystems west of 42°E) differs from the EEA’s EU-28. - Chemical recovery was confirmed quickly. - ICP Waters found sulphate falling in 11 of 12 regions of Europe and North America from 1990 to 2001, with widespread gains in alkalinity and pH (Skjelkvåle et al., 2005). - By 2003 there was “overwhelming documentation” of chemical recovery, but “little documentation of biological recovery” (Skjelkvåle et al., 2003). - Biological recovery was slow and incomplete. - At Lake Saudlandsvatn in southern Norway, deposition was five times the critical load. Chemical recovery began in the late 1990s. Brown trout recovered from 2003, but full recovery was not expected without further cuts (Hesthagen et al., 2011). - Rosseland (2021) describes recovery following “a pattern of hysteresis”: one critical acid episode can delay salmon recovery by 10–15 years. - Czech lake invertebrates only began to recover faster in the third decade after chemistry improved (Petruželová et al., 2023). - The concept’s later use. - Critical loads were extended to nitrogen. With integrated assessment modelling, they became the basis of the 1999 Gothenburg Protocol, the EU National Emission Ceilings Directive (2001) and the 2012 revision (Grennfelt et al., 2020; Hov and Grennfelt, 2026). - Insiders report that the idea began with “requests from both industry and negotiators” for a sounder basis for controls. Scientists were initially sceptical (Grennfelt et al., 2020). - They also note that the underlying models have been “criticised, not least from industry”, and credit their acceptance to transparency. - The limits of the approach. - For nitrogen, the same machinery has not closed the gap. The eutrophication critical load was still exceeded on 60.6% of the ecosystem area in 2024 (EMEP, 2026), against 84% in 1990 and 63% in 2010 (EEA). - Ammonia, overwhelmingly agricultural, fell only 17% in the EU from 2005 to 2023 (EEA, 2025).

Verdict: held up for sulphur and acidity. Exceedances fell well beyond the 60% gap-closure target, and lakes recovered chemically. Two qualifications: - biological recovery lags by decades; - the same approach has worked much less well for nitrogen, where abatement falls on a diffuse, politically protected sector.

Weight. Digest insight 8 (protecting a measurable intermediate threshold makes action tractable under uncertainty) is reinforced and can be graded moderate–strong. The hindsight adds a boundary condition. A threshold makes action tractable; it does not make it happen. Sulphur cuts came where the threshold coincided with fuel switching, desulphurisation and economic collapse. Nitrogen exceedances persist where abatement is costly for a diffuse sector.

Claim 9. By the late 1990s northern European SO2 emissions were down more than 50% from the 1980 maximum and UK urban concentrations more than 70%; lakes were in “slow, but steady” recovery and forest vitality improving, though “not unambiguous”; further cuts would follow under the 1994 protocol, and emphasis would shift to nitrogen, ammonia, VOCs, ozone and eutrophication (p. 106)#

Original claim. - Northern European emissions were down “more than 50 %” since 1980, and UK urban concentrations down “more than 70 %” (p. 106). - Lakes show “a slow, but steady, recovery”. “Even forest vitality is improving … although the evidence is not unambiguous” (p. 106). - The 1994 protocol “will lead to further SO2 emission reductions”. Emphasis is “shifting towards” interactions among nitrogen oxides, ammonia and volatile organic compounds (VOCs), combined acidification, photochemical oxidants and nitrogen eutrophication (p. 106).

What happened since. - Emissions: far beyond the forecast. - Europe (EMEP-reported): about 55 Tg (1980) to 15 Tg (2004). The European Gothenburg total for 2010 (16 Tg) was met by 2004 (Vestreng et al., 2007). - EU: SOx down 84% from 2005 to 2023, driven by energy, industry and transport measures (EEA indicator, 23 September 2025). - UK: down 98% since 1990, to 84 kt in 2024. The drivers were the switch from coal and fuel oil to gas and biomass, power-station closures and conversions, fuel-sulphur limits and desulphurisation (DEFRA, updated 12 February 2026). - Germany: 5.5 Mt (1990) to 0.2 Mt (2024), down 96%. The first-named driver is the closure or retrofitting of plants in the eastern Länder (Umweltbundesamt, updated 25 June 2026). - Shipping: the IMO global fuel-sulphur limit fell from 3.5% to 0.5% on 1 January 2020 (0.1% in emission control areas). The IMO forecasts a 77% cut in ship SOx (IMO). - Structural versus policy drivers. Crippa et al. (2016, Atmospheric Chemistry and Physics) estimate: - without technology and policy measures, EU SO2 emissions in 2010 would have been 129% higher; - with 1970 energy use but 2010 technology, they would have been 50% lower.

Both levers mattered. - Lakes: recovery confirmed, “steady” overstated. Chemical recovery is widespread and well documented (Claim 8). Biological recovery is slow, non-linear and incomplete (Rosseland, 2021; Hesthagen et al., 2011). - Forests: the improving trend did not hold. Mean defoliation has increased across Europe since 1990 (ICP Forests, 2025). German crown thinning is at or near its highest since 1984 (BMLEH, 2025). Sulphate deposition to forests fell by more than 60% over 2000–2020 (ICP Forests, 2025). The divergence confirms that crown condition was never mainly an SO2 indicator (Claim 7). - The shift of emphasis: correct. - The 1999 Gothenburg Protocol adopted exactly the multi-pollutant, multi-effect approach the chapter anticipated: SO2, NOx, ammonia and VOCs; acidification, eutrophication and ozone. - It was amended in 2012 and entered into force on 7 October 2019 (EMEP, 2026; Grennfelt et al., 2020). - Nitrogen and ozone are now the unresolved problems. The eutrophication critical load was exceeded on 60.6% of Europe’s ecosystem area in 2024. The WHO long-term ozone guideline was exceeded over nearly all of the EMEP domain (EMEP, 2026). - What the chapter did not foresee. - From around 2000, human health, above all fine particles, became the main driver of European air policy (Grennfelt et al., 2020; Fowler et al., 2020). - SO2 reductions also removed cooling sulphate aerosol, with an estimated +0.45 °C effect on European mean surface temperature from 1970–2010 measures (Turnock et al., 2016). Grennfelt et al. (2020) note “some trade-offs” between air-pollution and climate goals.

Verdict: partly held up. - The emissions and lake-chemistry claims, the forecast of further cuts and the forecast shift to nitrogen and ozone all held or were exceeded. - The forest-vitality claim did not hold. - The chapter missed the turn to health and the climate trade-off.

Weight. Strong support for digest insight 9 (large reductions come from structural and economic shifts as much as from regulation), now quantified. Use the forest sentence only as an example of how an indicator chosen during the alarm (crown condition) later tracked a different driver.

Claim 10. The EU’s 1999 SO2 limits (350 µg/m³ for 1 hour; 125 µg/m³ for 24 hours) are “not very much lower” than the smog-derived thresholds; 34% of Europe’s population exceeded the 24-hour limit in 1990 (pp. 102, 106)#

Original claim. - Lawther’s thresholds, derived from the London smog, were 500 µg/m³ for SO2 and 250 µg/m³ for black smoke (p. 102). - The EU’s 1-hour limit of 350 µg/m³ under Directive 1999/30/EC “is still not very much lower” than these (p. 106). - The 24-hour limit is 125 µg/m³. Progress “would seem” slow, as that limit was exceeded for 34% of Europe’s population in 1990 (Stanners and Bourdeau, 1995) (p. 106).

What happened since. - Compliance was achieved. The EEA’s exceedance indicator (published 30 April 2026) gives the share of the EU-27 urban population exposed above the SO2 daily limit (125 µg/m³) as “consistently below 0.1% since 2010”. In 2024, 0.2% was exposed above the WHO guideline level. This is not directly comparable with the chapter’s 34% (1990), which referred to Europe’s whole population and a different set of countries. Even so, the fall is more than two orders of magnitude. - Station-level exceedances. For 2023, the EEA reports exceedances of the EU daily SO2 limit only at 11 stations outside the EU (nine in Bosnia and Herzegovina, two in North Macedonia), which implies none within the EU. Concentrations above the WHO daily guideline were recorded in 13 of 33 reporting countries, 10 of them EU members (EEA, Europe’s air quality status 2024, 6 June 2024). - The limits were tightened, though the 1-hour value was not. Directive (EU) 2024/2881, Annex I: - Until 11 December 2026 the 1999 values continue: 350 µg/m³ for 1 hour (24 exceedances allowed a year) and 125 µg/m³ for a day (3 allowed). - From 1 January 2030: - 1 hour: 350 µg/m³, now with only 3 exceedances allowed; - 1 day: 50 µg/m³, 18 exceedances allowed; - a new annual limit of 20 µg/m³. - The alert threshold falls from 500 µg/m³ (1999/30/EC) to 350 µg/m³, with a new information threshold of 275 µg/m³. - Its recitals set out “a perspective for alignment” with the latest WHO guidelines by 2050 at the latest. That is an aim, not a binding limit. - WHO guidance moved in both directions. - The WHO’s 2021 global air quality guidelines set a 24-hour SO2 level of 40 µg/m³, defined as the 99th percentile. The 2005 guideline was 20 µg/m³. - Interim targets of 125 and 50 µg/m³ were kept. The 10-minute guideline of 500 µg/m³ remains valid. - The level rests on systematic reviews linking daily SO2 to mortality (high certainty) and asthma admissions (low certainty) (WHO, 2021; Orellano et al., 2021). - The EU’s 2030 daily limit therefore equals WHO interim target 2, and its old daily limit equals interim target 1. - Averaging times (now checked). The chapter compares a 1-hour limit (350) with Lawther’s threshold (500) but gives no averaging time for the threshold. The digest had already flagged the mismatch. - The WHO criteria document on sulphur oxides and particulates (Environmental Health Criteria 8, 1979) sets out what the chapter’s thresholds rest on. As lowest adverse-effect levels for short-term exposure, WHO chose “24-h mean concentrations” of 500 µg/m³ for SO2 and 500 µg/m³ for smoke, at which excess mortality might be expected. It chose 250 µg/m³ of each for worsening of patients with respiratory disease. - The same document reports Lawther’s own absolute-value analysis (1963): mortality increases became evident above 24-hour means of about 750 µg/m³ smoke and 710 µg/m³ SO2. - So the thresholds are 24-hour means. The chapter’s pairing (500 for SO2, 250 for smoke) mixes WHO’s two effect levels. Its source is WHO (1972), which I did not read, so the pairing may reflect that earlier document. - On a like-for-like basis, the 1999 daily limit (125) was a quarter of the 500 µg/m³ mortality threshold, and the 2030 daily limit (50) is a tenth. A 1-hour limit of 350 is also far stricter than a 24-hour mean of 500, because hourly peaks exceed the daily mean. - The 34% figure could not be checked. I could not access the relevant part of the 1995 Dobříš assessment.

Verdict: weakened. - The limit values the chapter quotes are correct. The 1-hour limit of 350 µg/m³ still stands for 2030. - The substantive claim, that limits are “not very much lower” than the smog-derived thresholds, does not survive a like-for-like comparison. The thresholds are 24-hour means (WHO, 1979), and against them the 1999 daily limit was already four times lower. - The “progress has been slow” judgement was overtaken. Exposure above the daily limit has been below 0.1% of the EU urban population since 2010, and the daily limit is being cut from 125 to 50 µg/m³.

Weight. Digest insight 12 (protective limits anchored on crisis thresholds can stay close to them for decades) should be graded weak and not used from this case. The averaging-time error is itself a small lesson: comparisons between limits need matching metrics. The WHO’s upward revision in 2021 also shows that evidence-based guideline values do not only ratchet down.

Minor factual checks#

Implications for the section’s transferable insights#

Stated in technology-neutral terms and keyed to the digest’s list.

Insight (digest #, pages) Effect of the post-2001 record
1. A fix that disperses harm improves the visible indicator while total harm grows (pp. 101–103, 107) Reinforced. Mylona’s data, and insiders’ accounts that tall stacks were seen as the solution in 1967, support it (Fowler et al., 2020; Grennfelt et al., 2020). Strong.
2. Positions on evidence follow who pays and who suffers (pp. 103, 107) Reinforced by primary records. UK ministers accepted the Scandinavian case in January 1985 but deferred action on value-for-money grounds, while emphasising others’ emissions (Hansard, 1985–86). Nordic insiders acknowledge that agendas often produced “outcomes that were favourable to them” (Hov and Grennfelt, 2026). Up from moderate to moderate–strong.
3. The boundaries of an appraisal decide its conclusion (pp. 103–104, 106) Strongly reinforced, differently. The decisive boundary turned out to be fine-particle health effects, worth 30:1 or more against costs. The EEA’s own 2013 volume reframes the case this way (Ch. 23, pp. 572–573). The chapter’s retrospective claim about materials remains uncited and unverified.
4. Economically trivial but visible harms carry public understanding (pp. 105–106) Supported. Insiders describe “acid rain” as a “bridging concept” and the forest alarms as “often exaggerated” (Grennfelt et al., 2020). Remains suggestive to moderate.
5. Action before causal consensus; overclaiming risks credibility (pp. 105, 107) Reinforced. Forest causation was never settled; the broad, proportionate justification was vindicated and the narrow one was not. Critics of precaution listed acid rain as a “false positive”, and the EEA (2013) had to rebut this. There is no sign of damage to policy support. Moderate–strong.
6. Long-term monitoring detects slow, diffuse change (p. 102) Strongly reinforced. The EMEP and ICP networks, running for 40+ years, documented recovery. The forest network founded in the scare now detects drought mortality.
7. Joint source–receptor knowledge turns blame into allocatable obligations (pp. 103–107) Reinforced. Blame matrices became routine and were extended to nitrogen and ozone. Strong.
8. Measurable intermediate thresholds make action tractable (pp. 106–107) Reinforced for sulphur: acidity exceedance fell from 43% to about 3.5%. Limited for nitrogen, where eutrophication exceedance is still about 60%. Tractable is not the same as delivered.
9. Reductions come from economic and structural shifts as much as regulation (pp. 104–106) Reinforced and quantified. The largest cuts came in 1990–99 in Central and Eastern Europe; German cuts came mainly from closures in the east. Technology and energy demand both mattered (Vestreng et al., 2007; Umweltbundesamt; Crippa et al., 2016).
10. Amplifying a warning ahead of the evidence sets the agenda (pp. 102–103) Supported. Odén’s article is described by insiders as “deliberatively provocative”. Sweden’s OECD delegate used intercontinental fallout from nuclear tests to win the argument (Grennfelt et al., 2020). Moderate.
11. Outsiders’ evidence counts once domestic institutions confirm it (p. 105) Partly supported. The UK set up its own review of domestic freshwater acidification (December 1984) before acknowledging the case. The order is not quite the chapter’s: the UK decided on desulphurisation in 1986, before the joint research programme reported in 1990. Suggestive.
12. Protective limits anchored on crisis thresholds persist (pp. 102, 106) Weak. The comparison mixes a 1-hour limit with 24-hour thresholds (WHO, 1979). Like for like, the 1999 daily limit was a quarter of the threshold. Daily and annual limits were tightened (2024/2881), and the WHO raised its daily guideline in 2021.

New lessons from hindsight, not drawn in the chapter (technology-neutral): - The main benefit of an intervention can come through a harm that was not the reason for acting. Fine-particle mortality, not fish or buildings, carried the economic case (Chestnut and Mills, 2005; Turnock et al., 2016). - Removing one harm can unmask another. Sulphate aerosol had been cooling Europe; removing it contributed to warming (Turnock et al., 2016). - Formal commitments can coexist with rising harm until an outside shock intervenes. The GDR and Czechoslovakia signed the 1985 protocol, but their cuts came only after 1990. - Chemical recovery precedes biological recovery by decades, with hysteresis. Soil damage persists after its cause is removed (Rosseland, 2021; Novotný et al., 2024).

Method and access notes#

Sources#

Primary historical, parliamentary and treaty records 1. Mylona, S. Sulphur dioxide emissions in Europe 1880–1991 and their effect on sulphur concentrations and depositions. Tellus B 1996;48(5):662–689, Table 1 (pp. 670–671). https://doi.org/10.3402/tellusb.v48i5.15939 ; PDF https://b.tellusjournals.se/articles/1249/files/submission/proof/1249-1-13694-1-10-20221115.pdf (accessed 25 Sep 2026) 2. Hansard, HC Deb 11 January 1985, vol 70, cc1011–73, “Acid Rain” (Waldegrave, cc1020–1023). https://api.parliament.uk/historic-hansard/commons/1985/jan/11/acid-rain 3. Hansard, HC Deb 18 December 1984, vol 70, c79W, “Acid Rain” (expert group on UK freshwater acidification announced 11 December 1984). https://api.parliament.uk/historic-hansard/written_answers/1984/dec/18/acid-rain 4. Hansard, HL Deb 29 October 1986, vol 481, cc709–11, “Acid Rain: EC Discussions”. https://api.parliament.uk/historic-hansard/lords/1986/oct/29/acid-rain-ec-discussions 5. Hansard, HC Deb 19 November 1986, vol 105, cc552–3, “Acid Rain” (Ridley: decision in principle on FGD announced 11 September 1986). https://api.parliament.uk/historic-hansard/commons/1986/nov/19/acid-rain 6. UN Treaty Collection. Protocol … on the Reduction of Sulphur Emissions … by at least 30 per cent, Helsinki, 8 July 1985 (status as at 25 Sep 2026). https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=XXVII-1-b&chapter=27&clang=_en 7. UN Treaty Collection. Protocol … on Further Reduction of Sulphur Emissions, Oslo, 14 June 1994 (status as at 25 Sep 2026). https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=XXVII-1-e&chapter=27&clang=_en 8. Council Directive 88/609/EEC of 24 November 1988 on large combustion plants (OJ L 336, 7.12.1988). https://eur-lex.europa.eu/eli/dir/1988/609/oj 9. Mason, B. J. (ed.). The Surface Waters Acidification Programme. Cambridge University Press, 1990 (online 1991). https://doi.org/10.1017/cbo9780511600067

London smog

  1. UK Met Office. The Great Smog of 1952 (case study; “about 4,000” known deaths; 750 in 1962). https://weather.metoffice.gov.uk/learn-about/weather/case-studies/great-smog (accessed 25 Sep 2026)
  2. Bell, M. L., Davis, D. L. Reassessment of the lethal London fog of 1952. Environ Health Perspect 2001;109(Suppl 3):389–94. https://doi.org/10.1289/ehp.01109s3389
  3. Bell, M. L., Davis, D. L., Fletcher, T. A retrospective assessment of mortality from the London smog episode of 1952: the role of influenza and pollution. Environ Health Perspect 2004;112:6–8. https://doi.org/10.1289/ehp.6539
  4. Davis, D. L., Bell, M. L., Fletcher, T. A look back at the London smog of 1952 and the half century since. Environ Health Perspect 2002;110:A734–5. https://doi.org/10.1289/ehp.110-a734
  5. Logan, W. P. D. Mortality in the London fog incident, 1952. Lancet 1953;1:336–8 (not read). https://doi.org/10.1016/s0140-6736(53)91012-5
  6. Scott, J. A. Fog and deaths in London, December 1952. Public Health Reports 1953;68(5):474–9 (report to the London County Council public health committee, 27 January 1953; weekly registered deaths, p. 476). https://stacks.cdc.gov/view/cdc/71913

Emissions, deposition and ecosystem status

  1. Vestreng, V. et al. Twenty-five years of continuous sulphur dioxide emission reduction in Europe. Atmos Chem Phys 2007;7:3663–81. https://acp.copernicus.org/articles/7/3663/2007/
  2. Fowler, D., Brimblecombe, P., Burrows, J., Heal, M. R. et al. A chronology of global air quality. Phil Trans R Soc A 2020;378:20190314. https://doi.org/10.1098/rsta.2019.0314
  3. Grennfelt, P., Engleryd, A., Forsius, M., Hov, Ø. et al. Acid rain and air pollution: 50 years of progress in environmental science and policy. Ambio 2020;49:849–64. https://doi.org/10.1007/s13280-019-01244-4
  4. Hov, Ø., Grennfelt, P. The legacy of the Nordic contributions to acid rain science and policy. Ambio 2026;55:1531–43. https://doi.org/10.1007/s13280-025-02304-8
  5. EMEP. Status Report 1/2026, Transboundary particulate matter, photo-oxidants, acidifying and eutrophying components (24 August 2026), Executive Summary and section 2.4.4. https://emep.int/publ/reports/2026/EMEP_Status_Report_1_2026.pdf
  6. EEA. Exposure of Europe’s ecosystems to acidification, eutrophication and ozone (CSI 005), assessment published 2 October 2018 (archived). http://web.archive.org/web/20241105023921/https://www.eea.europa.eu./data-and-maps/indicators/exposure-of-ecosystems-to-acidification-14/assessment-1
  7. EEA. Emissions of the main air pollutants in Europe (indicator, 23 September 2025). https://www.eea.europa.eu/en/analysis/indicators/emissions-of-the-main-air
  8. DEFRA. Emissions of air pollutants in the UK – Sulphur dioxide (SO2) (updated 12 February 2026). https://www.gov.uk/government/statistics/emissions-of-air-pollutants/emissions-of-air-pollutants-in-the-uk-sulphur-dioxide-so2
  9. Umweltbundesamt. Schwefeldioxid-Emissionen (updated 25 June 2026). https://www.umweltbundesamt.de/daten/umweltzustand-trends/luft/luftschadstoff-emissionen-in-deutschland/schwefeldioxid-emissionen
  10. International Maritime Organization. IMO 2020 – cutting sulphur oxide emissions. https://www.imo.org/en/MediaCentre/HotTopics/Pages/Sulphur-2020.aspx (accessed 25 Sep 2026)
  11. Crippa, M. et al. Forty years of improvements in European air quality: regional policy–industry interactions with global impacts. Atmos Chem Phys 2016;16:3825–41. https://doi.org/10.5194/acp-16-3825-2016

Freshwaters

  1. Skjelkvåle, B. L. et al. Regional scale evidence for improvements in surface water chemistry 1990–2001. Environ Pollut 2005;137:165–76. https://doi.org/10.1016/j.envpol.2004.12.023
  2. Skjelkvåle, B. L., Evans, C., Larssen, T., Hindar, A., Raddum, G. G. Recovery from acidification in European surface waters: a view to the future. Ambio 2003;32:170–5. https://doi.org/10.1579/0044-7447-32.3.170
  3. Hesthagen, T. et al. Chemical and biological recovery of Lake Saudlandsvatn … in response to decreased acid deposition. Sci Total Environ 2011;409:2908–16. https://doi.org/10.1016/j.scitotenv.2011.04.026
  4. Rosseland, B. O. The legacy from the 50 years of acid rain research. Ambio 2021;50:273–7. https://doi.org/10.1007/s13280-020-01408-7
  5. Petruželová, J. et al. Accelerated recovery of lake macroinvertebrates in the third decade since the reversal of acidification. Sci Total Environ 2023;892:164553. https://doi.org/10.1016/j.scitotenv.2023.164553

Forests

  1. Kandler, O., Innes, J. L. Air pollution and forest decline in Central Europe. Environ Pollut 1995;90:171–80 (pre-2001; not cited by the chapter). https://doi.org/10.1016/0269-7491(95)00006-d
  2. Pretzsch, H. et al. Forest stand growth dynamics in Central Europe have accelerated since 1870. Nat Commun 2014;5:4967. https://doi.org/10.1038/ncomms5967
  3. Ferretti, M. et al. (eds). On the pulse of European forests: 40 years of pan-European forest monitoring (ICP Forests 40th Anniversary Report). Thünen Institute, 2025. https://doi.org/10.3220/253-2025-21 ; PDF https://www.icp-forests.net/fileadmin/icp_forests/Dateien/Anniversary_Report/ICP_Forests_40th_Anniversary_Report.pdf
  4. BMLEH (German Federal Ministry of Agriculture). Ergebnisse der Waldzustandserhebung 2025 (Tables 2–3, 1984–2025). https://www.bmleh.de/SharedDocs/Downloads/DE/Broschueren/waldzustandserhebung-2025.pdf ; summary page https://www.bmleh.de/DE/themen/wald/wald-in-deutschland/waldzustandserhebung.html (accessed 25 Sep 2026)
  5. Hruška, J. et al. Forest growth responds more to air pollution than soil acidification. PLOS ONE 2023;18:e0256976. https://doi.org/10.1371/journal.pone.0256976
  6. Kolář, T. et al. Pollution control enhanced spruce growth in the “Black Triangle” near the Czech–Polish border. Sci Total Environ 2015;538:703–11. https://doi.org/10.1016/j.scitotenv.2015.08.105
  7. Novotný, R., Fadrhonsová, V., Šrámek, V. Long-term monitored Norway spruce plots in the Ore Mountains: 30 years of changes. Plants 2024;13:2379. https://doi.org/10.3390/plants13172379
  8. Metzger, B., Bemmann, M., Schäfer, R. Und ewig sterben die Wälder. Das deutsche Waldsterben als historisches Phänomen. Revue d’Allemagne 2007 (introduction read only). https://doi.org/10.3406/reval.2007.5967
  9. Metzger, B. “Erst stirbt der Wald, dann du!” Das Waldsterben als westdeutsches Politikum (1978–1986). Campus, 2015 (not read; review: Taschka, Central European History 2017, https://doi.org/10.1017/s0008938917000796)
  10. von Detten, R. (ed.). Das Waldsterben: Rückblick auf einen Ausnahmezustand. oekom, 2013 (publisher description only). https://doi.org/10.14512/9783865815682

Costs, benefits and materials

  1. Chestnut, L. G., Mills, D. M. A fresh look at the benefits and costs of the US acid rain program. J Environ Manage 2005;77:252–66. https://doi.org/10.1016/j.jenvman.2005.05.014
  2. US EPA. The Benefits and Costs of the Clean Air Act from 1990 to 2020 (Second Prospective Study, March/April 2011). https://www.epa.gov/clean-air-act-overview/benefits-and-costs-clean-air-act-1990-2020-second-prospective-study
  3. Turnock, S. T. et al. The impact of European legislative and technology measures to reduce air pollutants on air quality, human health and climate. Environ Res Lett 2016;11:024010. https://doi.org/10.1088/1748-9326/11/2/024010
  4. EEA. The costs to health and the environment from industrial air pollution in Europe – 2024 update (25 January 2024; updated 7 August 2026). https://www.eea.europa.eu/en/analysis/publications/the-costs-to-health-and-the-environment-from-industrial-air-pollution-in-europe-2024-update
  5. Tidblad, J., Kreislová, K., Faller, M. et al. ICP Materials trends in corrosion, soiling and air pollution (1987–2014). Materials 2017;10:969. https://doi.org/10.3390/ma10080969
  6. ApSimon, H. M., Cowell, D. The benefits of reduced damage to buildings from abatement of sulphur dioxide emissions. Energy Policy 1996;24(7):651–4 (abstract read; it reports no result). https://doi.org/10.1016/0301-4215(96)00058-4 ; abstract https://www.osti.gov/etdeweb/biblio/288718
  7. Cowell, D., ApSimon, H. Estimating the cost of damage to buildings by acidifying atmospheric pollution in Europe. Atmospheric Environment 1996;30(17):2959–68 (abstract read). https://www.osti.gov/etdeweb/biblio/288722

Air quality limits and guidelines

  1. Directive (EU) 2024/2881 on ambient air quality and cleaner air for Europe (recast), Annex I. https://eur-lex.europa.eu/eli/dir/2024/2881/oj
  2. Council Directive 1999/30/EC of 22 April 1999, Annex I (alert threshold 500 µg/m³). https://eur-lex.europa.eu/eli/dir/1999/30/oj
  3. WHO. WHO global air quality guidelines: particulate matter, ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide (2021), Table 0.1, section 3.6, Table 3.26. https://iris.who.int/handle/10665/345329
  4. Orellano, P., Reynoso, J., Quaranta, N. Short-term exposure to sulphur dioxide and all-cause and respiratory mortality: a systematic review and meta-analysis. Environ Int 2021;150:106434. https://doi.org/10.1016/j.envint.2021.106434
  5. EEA. Europe’s air quality status 2024 (6 June 2024). https://www.eea.europa.eu/en/analysis/publications/europes-air-quality-status-2024
  6. EEA. Exceedance of air quality standards in Europe (indicator, published 30 April 2026; data to 2024). https://www.eea.europa.eu/en/analysis/indicators/exceedance-of-air-quality-standards
  7. WHO. Environmental Health Criteria 8: Sulfur oxides and suspended particulate matter. Geneva: WHO, 1979 (section 1.1.8; Lawther 1963 discussed in the section on community exposures). https://www.inchem.org/documents/ehc/ehc/ehc008.htm

Framework and reception

  1. EEA. Late lessons from early warnings: science, precaution, innovation – Summary, EEA Report No 1/2013 (p. 24). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2
  2. EEA. Late lessons from early warnings: science, precaution, innovation, EEA Report No 1/2013, Annexes 1–3 (one file): Annex 2, Table A2.1 (p. 702) and Table A2.10 (p. 711); Annex 3 introduction and list of updated cases (pp. 717 ff.). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-annex-1/@@download/file
  3. Andersen, M. S., Clubb, D. O. Understanding and accounting for the costs of inaction. In EEA Report No 1/2013, Ch. 23, Box 23.2 (pp. 572–573). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-chapter-23/@@download/file
  4. Hansen, S. F., Tickner, J. A. The precautionary principle and false alarms — lessons learned. In EEA Report No 1/2013, Ch. 2 (p. 20; Table 2.3, p. 35). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-chapter-2/@@download/file
  5. Lidskog, R., Sundqvist, G. The role of science in environmental regimes: the case of LRTAP. Eur J Int Relat 2002;8:77–101 (abstract read). https://doi.org/10.1177/1354066102008001003
  6. Rothschild, R. E. Poisonous Skies: Acid Rain and the Globalization of Pollution. University of Chicago Press, 2019 (not read). https://doi.org/10.7208/chicago/9780226634852.001.0001