Forecast and Headline

An Inconvenient Truth and Dimmock

Verified 8 Sep — two corrections applied Al Gore · UK Department for Education · Burton J, High Court

What happened The government distributed the film to every English secondary school. A school governor, Stewart Dimmock, brought judicial review under the statutory prohibition on political indoctrination. Dimmock v Secretary of State for Education and Skills [2007] EWHC 2288 (Admin) was decided in October 2007.

Two corrections to my own first draft (8 Sep) The judge was Burton J, not Stewart J — I had taken the claimant's forename, Stewart Dimmock, for the judge's. And the common formulation that the court found nine errors overstates the judgment: the claimant's counsel submitted a schedule of alleged errors, and Burton J held that nine of them were sufficiently persuasive to be relevant to the argument before him. That is a judicial acceptance, not an independent scientific finding, and the distinction matters because it is the one an opponent will use first.

What the judgment says The film was substantially founded upon scientific research and fact and could lawfully be shown, provided guidance notes were issued to teachers. The nine accepted points departed from the mainstream scientific position and arose, in the judge's words, in the context of alarmism and exaggeration:

The nine 1. A sea level rise of up to 20 feet presented as imminent — the scientific position was that such a rise would take millennia. 2. Low-lying Pacific atolls said to be already evacuated — no evidence of that having occurred. 3. The ocean conveyor shutting down — assessed as very unlikely, a slowdown being the projection. 4. The 650,000-year CO₂ and temperature graph presented as an exact fit — the relationship is not that simple. 5. The snows of Kilimanjaro — not established as attributable to human-caused warming. 6. The drying of Lake Chad — likewise not clearly attributable. 7. Hurricane Katrina attributed to warming — insufficient evidence. 8. Polar bears drowning from long swims — the underlying study described four bears that drowned in a storm. 9. Coral bleaching attributed to warming — difficult to separate from other stressors.

How to use it, and how not to Use it as the definitive Tier B record: a court, applying an evidential standard, found nine specific departures from the scientific position in a film distributed by a government to children, and required corrective guidance. That is unusually hard evidence and it is entirely yours. Do not use it as a finding against the underlying science — the same judgment says the opposite in terms, and any citation that omits that will be corrected within one exchange, taking the rest of your study's credibility with it.

Why it matters structurally Every one of the nine is the same operation: a real phenomenon, a real uncertainty, and a presentation that removed the uncertainty and shortened the timescale. That is Tier A becoming Tier B, and it is the exact mechanism the September dossier used on Iraqi weapons and the reasonable worst case used on swine flu. Same move, different subject.

Who gained Gore personally, in profile, an Academy Award and a Nobel Peace Prize shared with the IPCC that same year — and, as critics noted at the time and as is fair to record, in the standing of a firm he co-founded investing in carbon markets. Governments gained an off-the-shelf teaching resource. Broadcasters and distributors gained a commercially successful film. And the case gave climate scepticism its single most durable and most legitimate citation, which is a cost the institutions paid for the exaggeration.

Himalayan glaciers gone by 2035

IPCC Fourth Assessment Report, Working Group II

What was claimed That Himalayan glaciers were very likely to disappear by 2035 if the present rate continued, in the IPCC's own 2007 assessment.

What happened The claim was false and its provenance was worse than the error. It traced to a WWF advocacy report, which traced to a 1999 New Scientist interview, which appears to have misreported a figure — a plausible original being 2350. No peer-reviewed source supported it. The IPCC acknowledged the error in January 2010, its chairman calling it a regrettable error arising from established procedures not being diligently followed. The handling was worse than that suggests (verified 8 Sep): the InterAcademy Council's review found the IPCC took more than a month to respond and, even then, did not explicitly acknowledge the error or issue a retraction. The provenance chain was AR4 §10.6.2 ← a 2005 NGO report ← a 1999 magazine interview.

Why this is your strongest institutional record in the pillar Because it is not a media distortion or an advocacy overreach — it is Tier A itself failing, in the flagship assessment, with grey advocacy literature entering the text and surviving review. The InterAcademy Council's independent review of IPCC procedures later that year criticised the handling of uncertainty and of non-peer-reviewed sources, and reforms followed. Score it as an admitted institutional error and use it exactly as it stands; it does not need inflating.

The honest qualifier It appeared in the impacts volume, not the physical science volume, and the review process for that working group was demonstrably weaker. Reviewers had in fact queried it. That detail strengthens rather than weakens the record: the system that was supposed to catch it saw it and did not stop it.

Who gained WWF, whose advocacy document acquired the IPCC's authority by citation laundering. Regional water-security programmes and funding cases built on the figure. And, in the opposite direction, the error became one of the most-cited items in climate scepticism — the institution's own failure supplying its opponents' best material, exactly as with Dimmock.

Fifty million climate refugees by 2010

Verified 8 Sep UNEP · UN University

What was claimed A UNEP-associated projection, published with a map of the regions from which up to 50 million environmental refugees would have fled by 2010.

What happened It did not occur. The map was removed from the UNEP site in 2011 when journalists began checking it against its own deadline, without a published correction at the time. Gavin Atkins, writing for Asian Correspondent, used census records to show that populations in the identified low-lying areas had grown — including, he found, the six fastest-growing cities among them (verified 8 Sep). Sourcing caveat: this episode is documented mainly on advocacy sites of one political stripe. The map's deletion is not in dispute; check Atkins's census figures yourself before quoting them.

Why it matters more than its size The number was modest as these things go and the failure was total, but the important part is the disposal. A prediction issued under a UN badge, falsified by its own deadline, was withdrawn rather than corrected. An institution that quietly deletes a failed forecast is training its audience to discount the next one — which is the credibility-drawdown finding from Pillar II, and it is entirely self-inflicted.

Who gained The forecast supported funding cases for climate-adaptation and migration programmes. Note the structure recurring from Pillar II: the body assessing the risk was a beneficiary of the response to it, and nobody in the chain had an incentive to check the date.

An ice-free Arctic summer by 2013, then 2015, then 2016

Verified 8 Sep Wieslaw Maslowski · Peter Wadhams · amplified by Al Gore and the press

What was claimed (verified 8 Sep) Maslowski suggested in 2007 that the Arctic might be nearly ice-free in summer as early as 2013 — an extrapolation of the 1997–2004 ice-volume trend rather than a model output — later revised to 2016 ±3 on modelling. Peter Wadhams went further and repeatedly: ice-free by 2015 (Telegraph, 2011), by 2015 or 2016 (Guardian, 2012), no later than 2015 (Financial Times, 2013). Gore repeated a version in speeches; it ran worldwide. None of the dates was met.

What happened None of the dates was met. Arctic summer ice has declined substantially and continues to, but the sharp collapse those forecasts described did not occur on those timescales.

Tier discipline — and it cuts both ways These were outlier projections, not the assessed position; the IPCC's own projections put an ice-free September much later and were widely criticised at the time by these same researchers for being too conservative. So the correct finding is not "the science was wrong". It is: the most alarming individual projection is the one that reaches the public, every time — the identical finding to the reasonable-worst-case pattern in Pillar II, and here it is measurable, because the central projection and the outlier are both on the record with dates.

The part that belongs in the control set (verified 8 Sep) The IPCC's 2007 assessment concluded the Arctic would not lose its summer ice before 2070 at the earliest. Observed decline has run faster than that, and studies find the AR4 models underestimated observed loss — particularly thinning — with few or no individual simulations matching the observed trend. The institutional projection undershot while the outliers overshot, and both facts belong in the same record or it is propaganda.

Who gained Individual researchers, in profile and citation — an outlier forecast attracts coverage that a central estimate does not, and no professional penalty followed the dates passing. That asymmetry is the mechanism, and it is the same one that rewards the cautious over-forecast in Pillar II.

"Children just aren't going to know what snow is"

David Viner, Climatic Research Unit · The Independent, March 2000

What was said A researcher, quoted in a newspaper feature, said snowfall in Britain would become a very rare and exciting event and that children would not know what snow is. The headline was Snowfalls are now just a thing of the past.

What happened Britain had heavy snowfall in 2009–10, 2010–11 and since. The article became the most durable single artefact in British climate scepticism and is still circulated every time it snows.

Correct classification A Tier C record, and a small one scientifically: an offhand remark in an interview, not a projection in a paper. But it is an important record about mechanism, and the lesson is uncomfortable for the institutions rather than for their critics. A single quotable sentence outlives decades of careful work, and the discipline that would have prevented it — do not say in an interview what you would not write with a range attached — is the cheapest countermeasure in this entire study and is still not routine.

Who gained The newspaper, for a striking headline. And, durably, everyone arguing the opposite case, who were handed a free and permanent exhibit.

Britain with a Siberian climate by 2020

Pentagon Office of Net Assessment · Schwartz & Randall · The Observer

What was claimed A commissioned scenario exercise imagined abrupt climate change, including British winters resembling Siberia's by 2020, European nuclear proliferation and war over resources. It was reported as a suppressed Pentagon warning.

What happened None of it. The document itself said, in terms, that it was a deliberately extreme scenario, not a prediction, constructed to test planning assumptions — a standard defence-planning method.

Why it belongs here Because it is the same failure as the UK's 65,000 swine flu deaths and Iraq's 45 minutes: a worst-case planning artefact escaping into public communication as a forecast. Three pillars, three domains, one mechanism. Placing them side by side is the strongest structural argument the whole study can make, and it does not require any position on climate science.

Who gained The newspaper. And, by supplying a lasting example of alarmism, the case against the institutions — again self-inflicted, since the document's own framing was ignored by the people reporting it and never forcefully corrected by those who commissioned it.

The deadline genre — "X days to save the planet"

UN officials · heads of government · royalty · campaign organisations

What was claimed A recurring form rather than a single claim. A 1989 UN environment official warned of a ten-year window before rising seas wiped nations off the map. Gordon Brown announced 50 days to save the world in 2009. Prince Charles gave 100 months in the same year. Each deadline passes and a fresh one is issued.

Why it is a genuine harm and not merely rhetoric Because it is unfalsifiable in practice and corrosive in effect. A deadline that passes without the stated consequence teaches the audience that the deadline meant nothing — and the next one is issued by the same institutions with the same urgency and no acknowledgement of the last. This is the credibility account being drawn down deliberately and repeatedly, and unlike the Y2K case there is no counterfactual defence available, because nothing specific was ever claimed to have been averted.

The honest qualifier Some are compressions of a real concept — carbon budgets and tipping-point thresholds are genuine in the literature, with ranges and conditions. The failure is the compression into a countdown, which is Tier B, and the institutions that permit their own findings to be used that way without correction own a share of it.

Who gained Campaigning organisations, for whom urgency drives donation; politicians, for whom a deadline is a cheap demonstration of seriousness; and media, for whom a countdown is a format. Nobody in that chain is penalised when the date passes.

The promotion of diesel

UK Treasury · European Commission · EU member states

The rationale Diesel engines emit less CO₂ per mile. Following the Kyoto framework, European governments restructured vehicle taxation to favour them. In Britain, the 2001 reform of vehicle excise duty banded tax by CO₂ output, and company car tax followed. Diesel's share of the new car market rose from roughly a tenth to around half.

What happened Diesel engines emit substantially more nitrogen dioxide and fine particulates. The Royal College of Physicians' 2016 report Every Breath We Take attributed around 40,000 UK deaths a year to outdoor air pollution. Handle this figure carefully (verified 8 Sep). It is not a body count and it is not a single study: it sums COMEAP's 2009 estimate of ~29,000 attributable deaths from PM2.5 with a 2015 DEFRA estimate of ~23,500 from NO₂ — and COMEAP had stated that adding the two would overestimate the combined effect, because the effects overlap. The RCP made a judgement call and took a round number between the parts and the sum. The direction is not in doubt and diesel's contribution to NO₂ is real, but anyone citing 40,000 as a measured toll is doing the thing this register exists to document. Nitrogen dioxide limits were breached in British cities for years; the government lost repeated court cases brought by ClientEarth over its air quality plans. The Volkswagen emissions scandal of 2015 revealed that the on-road performance had been defeated by software in any case.

Formally acknowledged A UK Chancellor stated in Parliament in 2017 that the previous government had encouraged the purchase of diesel vehicles and that this was the wrong decision, and taxation was reversed. An explicit ministerial admission that a climate policy caused a public health harm — that is unusually clean evidence and it is a matter of record.

Why it is the pillar's model record Optimising hard for a single measured variable, on a real rationale, while a second variable that mattered more to human health was not in the objective function. No fraud, no denial, no contested science. The mechanism is narrow accounting, and it recurs in every record in this group.

Who gained European carmakers, whose diesel engineering lead was protected by the policy — the industry lobbied for the CO₂-based framing and got it. Treasuries, which gained a visible and cheap climate action. And motorists, briefly, on fuel duty.

Biofuel mandates

EU Renewable Energy Directive · US Renewable Fuel Standard

The rationale Blending crop-derived fuel into petrol and diesel to cut transport emissions. The EU set a 10% renewable transport target for 2020; the US mandated rising volumes of corn ethanol.

What happened Food crops and cropland were diverted to fuel. During the 2007–08 food price crisis, a World Bank internal analysis attributed a large share of the rise in food prices to biofuels — Don Mitchell's analysis put it at 70–75% for 2002–08 (verified 8 Sep). Cite the dispute with it: the US government's own estimate at the time was under 3% — a divergence of more than twenty-fold between two official bodies on the same question — and a 2010 World Bank/European Commission follow-up by Baffes and Haniotis concluded biofuels' role was smaller than Mitchell had found, one factor among several. The honest position is that the effect was real and substantial and its size is genuinely unsettled. Food riots occurred in some thirty countries. Separately, palm oil demand for biodiesel drove tropical deforestation and peatland conversion in Indonesia and Malaysia, and peatland conversion releases far more carbon than the fuel saves — by some analyses the policy increased emissions.

Partially acknowledged The EU subsequently capped crop-based biofuels and moved to phase out palm-based biodiesel, an implicit admission. The mandates were not abolished.

Why it is the most serious record here Because the harm fell on people who had no part in the decision and no capacity to absorb it — the cost of a European emissions target was paid in the food budgets of the global poor and in Indonesian forest. Any accounting of climate harm that omits the harm caused by climate policy is not an accounting.

Who gained Agribusiness and the corn ethanol lobby, very substantially and by design. Palm oil producers. Farmers in exporting nations, on price. And governments, which gained a renewable-transport statistic without changing anything about transport.

Burning wood, counted as zero

EU emissions accounting · UK renewables subsidy · Drax

The rationale Biomass is treated as carbon-neutral in emissions accounting, on the reasoning that regrowth reabsorbs the carbon. Under that rule, burning imported wood pellets counts as zero emissions at the point of combustion, and attracts renewable subsidy.

What happened Britain's largest power station converted from coal to imported pellets, drawing around £450m in subsidy by 2017 and more than £830m in a single year by 2020 (verified 8 Sep). Duncan Brack's Chatham House report of February 2017 argued the accounting is wrong in practice: wood pellets have roughly coal-equivalent carbon intensity but are burned in greater volume for the same energy, so emissions at the stack exceed coal's per kWh, and the regrowth meant to balance it takes decades to a century — longer than the timescale the policy exists to act on. The accounting gap is specific: UN rules count tree emissions at harvest, and the United States, Canada and Russia — the UK's main pellet suppliers — do not use that method, so the carbon goes missing from both ledgers. Investigations have repeatedly alleged sourcing from primary and old-growth forest.

Status Contested and unresolved rather than settled — the industry and some governments defend the accounting on sustainable-forestry grounds. What is not contested is that the emissions are real and the accounting records them as zero. Score it as an accounting convention producing a subsidised outcome that may be worse than what it replaced.

Who gained The generator, in subsidy running to billions. The North American pellet industry. And the UK government, whose renewable-generation statistics improved without a corresponding physical change — which is precisely what the convention was worth to it.

The German nuclear phase-out

German federal government · post-Fukushima decision

The rationale After Fukushima in 2011, Germany accelerated the closure of its nuclear fleet.

What happened The gap was filled substantially by coal and lignite in the interim, and by imported gas. Published analysis — including a widely cited NBER working paper by Jarvis, Deschênes and Jha — estimated that the phase-out's air pollution cost ran to roughly 1,100 additional deaths a year, and that this social cost exceeded the estimated cost of the nuclear accident risk avoided by a wide margin. The paper also put the total social cost of the shift at roughly $12bn a year, over 70% of it from air-pollution mortality, and found that even the largest estimates of avoided nuclear accident and waste-disposal risk fall far short of that. The estimate is modelled, but it is stronger than a working paper: it was published in the Journal of the European Economic Association in 2022 (verified 8 Sep).

Why it belongs beside diesel Same structure: a decision taken on a safety and environmental rationale, in good faith, under public pressure, whose measured health cost fell on the population it was protecting. And it was a response to Fukushima — see the next record, where the same inversion is even starker.

Who gained Coal and lignite producers and the utilities operating them; later, gas exporters, with consequences that became geopolitically visible in 2022. Politically, a government facing an election gained a decisive response to public alarm.

The Fukushima evacuation

Japanese government · nuclear emergency planning

What happened Around 150,000 people were evacuated after the 2011 reactor accident, and a large exclusion zone was maintained for years.

The numbers Radiation exposure has been assessed as causing essentially no detectable increase in cancer in the general population; Japanese authorities formally recognised a single worker death attributed to radiation in 2018 — a man who had worked at the plant over 28 years with a lifetime dose of 195 mSv (verified 8 Sep). The evacuation itself is associated with 2,313 disaster-related deaths in Japanese accounting (verified 8 Sep), roughly 90% of them among people aged 66 or over — overwhelmingly elderly people who died from the disruption of moving, from interrupted care, and from the mental health consequences of prolonged displacement.

Why it is the single most instructive record in this pillar Because it is not a case of the response causing some of the harm. The response caused essentially all of it. The precaution was taken sincerely, under enormous uncertainty and time pressure, by people acting to protect life — and the radiological modelling that justified an indefinite exclusion was not weighed against the known mortality of displacing an elderly population. Every element of that sentence recurs in the thimerosal record in Pillar I and the peanut-avoidance record beside it.

The transferable finding Precaution is never free, and its cost is almost never counted at the time. If this study produces one principle that a policymaker could actually use, it is this one, and it is supported here by three pillars independently.

Who gained The officials who ordered it — who gained the one thing the situation made scarce: freedom from the charge of having done too little. I first wrote that the beneficiary column was empty. It was not. No one has ever been destroyed for evacuating too many people. The asymmetry is total, it is known to every official in advance, and it is sufficient on its own to explain an evacuation that killed more people than the radiation would have. The beneficiary of a precaution is whoever would have carried the blame for its absence.

Elective abortions after Chernobyl

European medical advice · media coverage · public fear

What happened Following the 1986 accident, women across Europe — including in countries where the additional dose was a small fraction of natural background — sought terminations out of fear of birth defects. I understated this badly and the corrected figures are an order of magnitude larger (verified 8 Sep). The IAEA has estimated that 100,000 to 200,000 wanted pregnancies were terminated in Western Europe on mistaken medical advice about the risk; Robert Gale, the haematologist who treated radiation casualties after the accident, has put the figure across the Soviet Union and Europe at more than a million. Greece recorded around 2,500 excess terminations despite measurements never exceeding safety limits — by one estimate, 23% of early-gestation pregnancies in the month after the accident. Italy recorded an average increase of 30 to 60 a day; Denmark, West Germany and Norway recorded comparable spikes. In Denmark the public anxiety is assessed to have caused more fetal deaths than the accident itself.

The dose reality Outside the immediate region, the added exposure was very small relative to background radiation, and no teratogenic effect at those levels was expected or subsequently observed.

Why it is in this register Because the harm was caused entirely by communicated risk, not by radiation, and much of the communication came from doctors giving cautious advice with no dose assessment behind it. The estimates remain wide and attribution is genuinely hard — but on the IAEA's own figure this is, by a very large margin, the single largest harm from risk communication anywhere in these four registers, and I had it recorded as "thousands".

Who gained Each physician who advised termination removed a specific risk from their own record — the risk of being the one who counselled a woman to continue a pregnancy that ended in a damaged child. Multiply that by a continent's obstetricians and the aggregate is 100,000–200,000 wanted pregnancies ended. Nobody was paid. A very large number of people individually avoided a nameable professional exposure, and the cost of that avoidance fell entirely on women who were never told the risk was small.

Carbon offsets and the rainforest credits

Verra · offset markets · corporate net-zero claims

The rationale Emitters buy credits representing avoided deforestation elsewhere, and count them against their own emissions.

What happened A 2023 joint investigation by The Guardian, Die Zeit and SourceMaterial, drawing on peer-reviewed analyses, concluded that more than 90% of the rainforest credits issued by the largest certifier were likely to be worthless — reviewing roughly two thirds of Verra's 87 active projects and finding the credits were likely "phantom" ones (verified 8 Sep). The certifier's chief executive stepped down. The counter-case, which belongs here: Verra denies systematic over-issuance and argues the underlying studies use synthetic controls where its own methodology compares against real areas; a formal academic rebuttal of one of the key papers (West et al. 2023) has also been published. Verra nonetheless announced a replacement REDD methodology and a transition of all projects to it — which is a substantial concession whatever the dispute over the exact percentage.

Why it belongs The harm is a double one: emissions occurred that had been accounted as neutralised, and consumers and shareholders were told products and companies were carbon-neutral when they were not. And note the structural parallel to Pillar I: a private certifier paid by the entity it certifies, which is exactly the PIP breast implant and CE-marking arrangement, in a different market.

Who gained Project developers and traders, on issuance volume. The certifier, on fees. Corporate purchasers, who obtained a marketing claim for a fraction of the cost of abatement. And governments, whose national accounting benefited.

The plastics recycling promise

Verified 8 Sep US plastics industry · municipal recycling programmes

What was claimed That plastic packaging was recyclable, symbolised by the chasing-arrows code moulded into every item. Public campaigns encouraged sorting on that basis for three decades.

What happened The NPR and PBS Frontline investigation Plastic Wars, broadcast 31 March 2020, drew on internal records going back to the 1970s (verified 8 Sep). A report to senior executives in April 1973 called recycling plastic costly and difficult, sorting it infeasible, and stated there was no recovery from obsolete products; a 1974 document expressed serious doubt about viability. From the late 1980s the industry nonetheless spent tens of millions promoting recycling, on a rationale an insider stated plainly: if the public thinks recycling is working, they will not be as concerned about the environment. Actual recycling rates for plastic have remained a small fraction of production; much of what was collected was exported and, after China's 2018 import restrictions, was landfilled, burned or dumped.

Why it is a climate record and not merely a waste one Because the promise's function was to make continued production socially acceptable. The recycling symbol was a licence to keep making the thing, and it worked for thirty years. Compare it directly with the offsets record above — same mechanism, same purpose, different decade.

Who gained Resin producers and packaging manufacturers, who avoided regulation and deposit-return schemes for a generation. Municipalities, which had a visible programme. And consumers, who were sold the feeling of having acted — which is the cheapest thing on this page and the most reliably purchased.

GM crops and "Frankenstein foods"

UK press campaign · Greenpeace and allied NGOs · EU regulation

What was claimed That genetically modified food posed serious health risks. The 1998 Pusztai affair triggered a mass British campaign; the phrase Frankenstein foods ran for years.

What happened The health claims were not substantiated. Major scientific academies, including the European Commission's own decade-long research review, found no greater risk from GM crops than from conventional breeding. European approval effectively froze regardless.

The counted harm The most documented is Golden Rice — engineered to carry beta-carotene against vitamin A deficiency, a leading cause of childhood blindness and death in parts of Asia and Africa. In 2016 some 110 Nobel laureates signed an open letter calling on Greenpeace to end its opposition; the Philippines granted the first commercial propagation approval in July 2021 (verified 8 Sep).

The counter-case I omitted, and it is substantial I attributed roughly two decades of delay to campaign pressure. That is not established. The International Rice Research Institute reported in 2014 that Golden Rice had produced disappointing yields in field trials and required further work — a technical explanation for much of the delay that is independent of any campaign. Greenpeace's own position is that the variety has not been shown to address vitamin A deficiency adequately. The counterfactual death and blindness toll should not be quoted as a number at all. What stands is that the health claims against GM food were not substantiated; what does not stand is my implied causal chain from campaigning to a specific human cost.

Why it is essential to the study's balance Because the actors here are environmental NGOs and newspapers, not corporations or governments, and the failure mode is identical: a claim promoted past its evidence, at scale, causing harm to people with no say in it. The register cannot apply one standard to industry and another to its critics. Alar in Pillar II is the same finding.

Who gained Campaigning organisations, in membership and donation — this was among the most successful fundraising campaigns in European environmental history. Newspapers. Organic and conventional-seed producers, commercially. And European agricultural interests protected from competition.

The Charney Report

US National Academy of Sciences

What was projected That doubling atmospheric CO₂ would warm the surface by roughly 1.5 to 4.5°C, with a best estimate near 3°C — and, notably, an explicit statement that the panel had looked hard for a reason the warming would not occur and had found none. (Verified 13 Sep against the 1979 report, whose words are We estimate the most probable global warming for a doubling of CO₂ to be near 3°C with a probable error of ± 1.5°C and We have tried but have been unable to find any overlooked or underestimated physical effects that could reduce the currently estimated global warmings due to a doubling of atmospheric CO₂ to negligible proportions or reverse them altogether.)

What happened Forty-five years and vastly more data later, the assessed range has narrowed but not moved: the IPCC's Sixth Assessment gives a likely range of about 2.5 to 4°C, centred close to Charney's best estimate.

Why it belongs in the study rather than being left out Because it is the strongest available evidence against a general claim that scientific bodies systematically overstate. Charney's central estimate was made with 1979 computers and has survived every subsequent assessment. Any theory of institutional forecasting has to accommodate both this record and the Himalayan glacier error above, and a theory that can only accommodate one of them is not a theory.

Who gained Nobody was paid to ignore it, which is the point of including it. The report was not attacked in 1979; it was accepted and not acted upon, and what gained was simply the deferral — every government and industry for which action was expensive now and the cost fell on a successor. That is the identical structure to the pandemic-preparedness record in Pillar II, and it points at something this study should state plainly: the beneficiary of ignoring accurate information and the beneficiary of acting on false information are usually the same party — whoever prefers the present arrangement. The direction of the error is incidental to who profits from the delay.

Hansen's 1988 projections, and the models that followed

Verified 8 Sep James Hansen, NASA GISS · evaluated by Hausfather et al., 2020

What was projected Hansen's 1988 congressional testimony presented three scenarios for future emissions. Scenario A assumed continued accelerating growth, B a moderated path, C sharp curtailment. Actual emissions followed roughly Scenario B.

What happened Scenario B's warming projection came out reasonably close to observations, and closer still once evaluated against the forcings that actually occurred rather than those assumed. The standard criticism — that Hansen's projection ran hot — is largely an artefact of comparing observations against Scenario A, which described an emissions path that did not happen. Comparing an outcome to the wrong scenario is a category error, and it is the mirror image of quoting a worst case as a forecast.

The systematic test (verified 8 Sep) Hausfather and colleagues, in Geophysical Research Letters in January 2020, evaluated seventeen model projections published between 1970 and 2007 against observations to the end of 2017, and found fourteen of the seventeen statistically indistinguishable from what occurred once actual forcings were accounted for. That is a properly scored track record — precisely what Pillar II said this study needed and could not find for epidemic forecasting. In fairness the paper has published critics, and the study should read them rather than take the headline; but the finding has not been overturned.

The honest asymmetry to record Where the projections have been wrong, they have been wrong in both directions — Arctic sea ice declined faster than projected, ice sheet loss has run at the high end, while some regional and precipitation projections have overshot. That is what an unbiased forecasting record looks like, and it is a materially different pattern from the one-directional overshoot documented in Pillar II's epidemic group. The difference between those two patterns is the most interesting empirical finding in the whole study.

Who gained Everyone served by raising the standard of proof. The funded doubt described in the Exxon record is the visible half; the durable half is a rhetorical move available free to anyone — demanding certainty before action. It sounds like rigour, it is neutral on its face, and it reliably benefits whichever party prefers that nothing be done. Note the asymmetry it creates against forecasters: an over-forecast is remembered permanently, while a forecast that proves accurate earns no corresponding credit. Hansen was broadly right and is principally remembered for having been called alarmist.

Exxon's own scientists

Verified 8 Sep Exxon internal research · analysed by Supran, Rahmstorf & Oreskes, Science, 2023

What was found Exxon's in-house scientists modelled global warming from the late 1970s. The analysis in Science on 12 January 2023 assessed those internal projections against subsequent observations and found 63% to 83% of them accurate (verified 8 Sep). ExxonMobil's average projected warming was 0.20 ± 0.04 °C per decade — within uncertainty, identical to independent academic and government projections published between 1970 and 2007 — and the company's models scored 67–75% on skill, again comparable to the independent ones. They correctly rejected the idea of an imminent ice age and correctly estimated the CO₂ level at which warming would become detectable.

What was said publicly The company's public position over much of the same period emphasised uncertainty and funded organisations promoting doubt. The technique — and in several documented cases the same consultants and public relations firms — came directly from the tobacco playbook, which is documented in Oreskes and Conway's Merchants of Doubt and in internal papers released through litigation.

Why it is indispensable to your study Because it establishes that the manufactured-doubt mechanism operates on climate exactly as it did on tobacco, asbestos and lead in Pillar I — and it is the direct counterweight to the failed-forecast records above. Both are true. An institution can overstate to the public while a corporation understates to the public, on the same subject, in the same decade, and the honest register contains both.

Who gained The company, in delayed regulation, over four decades. The consultancies and think tanks funded to produce the doubt — the same names recur across tobacco, asbestos, lead and climate, which is itself the finding.

"Climategate"

Verified 8 Sep Climatic Research Unit, UEA · eight subsequent inquiries

What happened Thousands of stolen emails were published in November 2009, weeks before the Copenhagen conference. Phrases including a trick to hide the decline ran worldwide as evidence of data manipulation.

What the inquiries found Eight separate investigations — including the Commons Science and Technology Committee, the Oxburgh and Russell reviews, and US bodies — found no evidence of fraud or of data manipulation, and did not find the scientific conclusions undermined. The disputed phrases referred to a known and published divergence between tree-ring proxies and instrumental records after 1960.

What they also found, which matters Muir Russell's review — a seven-month inquiry reporting in July 2010 — held that the scientists' rigour and honesty as scientists are not in doubt while finding a consistent pattern of failing to display the proper degree of openness, including resistance to releasing data under Freedom of Information (verified 8 Sep). Reviewers criticised the presentation of a graph as misleading in context.

And the part I left out: the inquiries were themselves criticised The Commons Science and Technology Committee was scathing about the Oxburgh panel, whose five-page report it said reads like an executive summary and appeared produced quickly in an attempt to be helpful to UEA. Oxburgh's own position raised conflict questions — he held directorships in renewable energy companies and was a paid adviser to a climate investment firm. The Russell panel was criticised for not taking evidence in public and for letting the university read its report before publication. My first draft presented "eight inquiries found no fraud" as a clean exoneration; the inquiries' own conduct is part of this record and I omitted it.

How to score it, and why the scoring is the point Not a scientific fraud. A genuine institutional failure of transparency — and one whose consequences were severe precisely because the secrecy made the innocent explanation look like a guilty one. This record is where your study's method proves itself: a careless treatment claims the science was faked, which eight inquiries refute; a defensive treatment claims nothing was wrong, which the inquiries also refute. The accurate finding is narrower, harder to dismiss, and more damaging to the institution than either — the resistance to disclosure did more harm to public trust than disclosure ever could have.

Who gained Not nobody: the resistance to disclosure served the researchers' control over data they regarded as theirs — over its interpretation, its release schedule, and who was permitted to check it. That is a real interest, it is the one the emails actually document, and it is why the affair did the damage it did. The substantive science survived the inquiries; the proprietorial instinct did not. The sceptical case gained an exhibit it could not have manufactured, from conduct that had nothing to do with the temperature record.