Late Lessons, Jensen Huang and AI
Markdown version

LL1-10 digest: Ch10 Sulphur dioxide: from protection of human lungs to remote lake restoration#

Late lessons from early warnings (EEA, 2001), pp. 101–109. Author: Arne Semb, a chemist at the Norwegian Institute for Air Research (NILU). He took part in the OECD long-range transport programme, the Norwegian SNSF project and EMEP (p. 198). The chapter has no panels. Table 10.1 is credited “Source: EEA”, which suggests the editors compiled it.

Core story#

After the London smog of December 1952 (more than 2,000 deaths in a week, p. 101), Britain went after smoke, not sulphur. London depended on coal, and coal sulphur could not easily be removed. Under a “best practicable means” regime with no emission limits, the main response to sulphur dioxide (SO2) was to build taller chimneys “in proportion to the emitted amounts” (p. 101).

Ground-level air got cleaner while total European emissions more than doubled after 1950, peaking at about 57 Mt in the mid-1970s (Fig. 10.1, p. 103). The electricity industry was “confident” that dilution made emissions harmless. “This was not so” (p. 102).

The warning came from Sweden. A precipitation-chemistry network built to study plant nutrients showed that rain was acidifying. Svante Odén announced this with “sweeping statements”, first in a newspaper article (dated 1967 in the references) and then in a 1968 report. Sweden’s “creative presentation” at the 1972 Stockholm conference worked “as intended” (p. 102). Semb reports this without criticism. Two research programmes followed. The OECD study (1972–77) linked Scandinavian acidification “quantitatively” to other countries’ emissions (p. 103). Norway’s SNSF project documented falling trout and salmon stocks by 1976; forest effects were “much more difficult to demonstrate” (p. 103).

Resistance stayed strong. A 1977 Nature editorial called acid rain “a million-dollar problem with a billion-dollar solution”. A Norwegian official replied that emitters were ignoring their own damage: “Europe versus itself” (pp. 103–104). An OECD study in 1981 found fish and timber losses small, but damage to buildings comparable to the cost of abatement (p. 104).

Cold War diplomacy produced the EMEP monitoring programme (1976) and the transboundary air pollution convention, CLRTAP (1979) (p. 104). West German alarm over forest decline, fanned by Der Spiegel and, on Semb’s account, Green Party pressure, pushed Germany to act without agreed causes (pp. 104–105, 107). The 1985 protocol set a 30% cut; the UK and Poland did not sign (p. 105). A joint research programme of the UK Royal Society and the Norwegian and Swedish academies, which Semb calls “somewhat superfluous”, confirmed that acid rain killed fish. The UK generating board’s chairman then had to explain to the Prime Minister that Norway’s charges were “scientifically well founded” (p. 105; undated, the editors put UK acknowledgement at 1985, p. 180).

Without a proven dose–response for forests, the convention’s effects task force developed “critical loads”: the most deposition an ecosystem can take without unacceptable change. Critical loads were mapped across Europe and became the basis of the 1994 protocol, although only “60 % gap closure” was found achievable at the worst sites (pp. 106–107).

By the late 1990s, northern European emissions were down more than 50% from their peak and UK urban concentrations more than 70%. Lakes were recovering slowly; forest recovery was “not unambiguous” (p. 106).

Semb’s own verdict (pp. 106–107)#

“Generally” action came only on “proof beyond reasonable doubt” (p. 106). “Only when the issue was taken to the international level could significant change occur.” Precaution was “altogether absent” early on, “mainly because decision-makers had limited understanding”. Tall stacks showed “reluctance to accept effects beyond the immediately obvious” (p. 107).

Positions followed perceived costs: Scandinavia found action “relatively easy” to argue for, the UK moved “arguably” only when costs “closer to home” became clear, and Eastern Europe could not pay. What changed things was the ECE convention (also a tool of East–West détente) and German Green Party pressure. The German advisers’ reasoning, which Semb labels a precautionary approach, was “for the record”: “Politically, the decision had already been taken.” Critical loads shifted the burden of proof from monetising future damage to finding a no-chemical-change deposition level, a change in what must be shown rather than who must show it: “not exactly” precaution, but “a rational way of dealing with uncertainties” (p. 107). There are no explicit recommendations.

Main mechanisms#

Dispersion instead of reduction. Positions (and possibly acceptance of evidence) following where a country sat. Contested boundaries of cost accounting. Symbolic harms carrying the case. Politics acting before causal consensus. Shared monitoring giving a common basis for allocating responsibility (necessary, not sufficient). Critical-load reframing. Commercial alignment and outside shocks (refinery profits, the oil crisis, gas, and the Eastern collapse as a “new situation” negotiators exploited; the chapter does not say the collapse itself cut emissions). Détente sustaining the institutions. Dependence on one fuel limiting remedies.

Transferable insights (technology-neutral)#

  1. A fix that disperses or relocates harm can improve the visible indicator while total harm grows and shifts to distant receivers (pp. 101–103, 105, 107). Strong within this case.
  2. With cross-border harms, willingness to act follows who pays and who suffers (pp. 103, 107); that acceptance of evidence does too is the editors’ hedged suggestion (p. 180). Moderate for action, suggestive for evidence. Interests are inferred, not documented.
  3. The boundaries of an economic appraisal can decide its conclusion; counting diffuse home-country damage changes the verdict (materials damage alone was “comparable” to abatement costs, p. 104) (pp. 103–104, 106). Moderate. The stronger claim, that costs were “more than” recovered, is uncited.
  4. Economically trivial but visible harms can carry public understanding beyond the science (pp. 105–106). Suggestive.
  5. Action can come before causal consensus. Experts then warn that overclaiming a cause risks credibility, and prefer proportionate, broad justification (pp. 105, 107). Moderate.
  6. Long-term monitoring, even when built for other purposes, detects slow, diffuse change (p. 102). Moderate.
  7. Jointly produced source–receptor knowledge gives a common basis for allocating obligations (pp. 103–107). Moderate. It was necessary but not sufficient: resistance continued for years after the 1977 attribution (pp. 103, 105).
  8. When causes stay unknown, protecting a measurable intermediate threshold with safety margins makes action tractable, at the price of simplification and compromise (pp. 106–107). Moderate.
  9. Big reductions can be enabled by independent economic and political shifts alongside regulation, and the chapter does not separate their contributions (pp. 104–106). Moderate. An industry whose profits align with abatement may not resist it (one uncited sentence, p. 104). Suggestive.
  10. Vivid communication of a warning ahead of full documentation can set the agenda; years of documentation work followed (pp. 102–103). Suggestive. Semb does not say the amplification caused the delay.
  11. Outsiders’ evidence may count only once research involving the sceptic’s own institutions confirms it (here a joint UK–Norwegian–Swedish programme) (p. 105; editors p. 180). Suggestive.
  12. Protective limits anchored on crisis thresholds can stay close to them for decades (pp. 102, 106). Suggestive. The averaging times do not match, and the “slow progress” figure (1990) predates the 1999 limit.

Main caveats#