Late Lessons, Jensen Huang and AI
Markdown version

LL1-11 hindsight check: Ch11 MTBE in petrol as a substitute for lead#

Late lessons from early warnings: the precautionary principle 1896–2000 (EEA Environmental Issue Report No 22, 2001), pp. 110–125. Authors: Martin Krayer von Krauss and Poul Harremoës. Editing finished 10 May 2001 (p. 110).

Checked on 25 September 2026 against sources published up to that date.

Method and limits. - I used primary sources wherever I could reach them: the Federal Register, US public law, court opinions, SEC filings, IARC, ATSDR, US EPA, USGS, the California Water Boards, Anses opinions under REACH, EUR-Lex and WHO. Peer-reviewed papers came next. - Some sources are industry-affiliated or industry-funded. They are flagged where used: Concawe, API, GSI Environmental, and the Borghoff et al. and Rivera et al. reviews. - Several primary pages were blocked to automated access. For those I relied on indexed excerpts or on secondary descriptions, and say so in the text: - ECHA’s substance pages; - EUR-Lex full text of Recommendation 2001/838/EC; - the final ATSDR 2023 profile PDF (I read the January 2022 public-comment draft instead); - Springer abstracts. - Quotations from the Anses opinions are my own translations from the French. - My web-search allowance ran out near the end of the work, so three points could not be verified further: - the effectiveness of EU underground-tank enforcement; - the published version of Arvin’s sand-filter work; - MTBE’s listing status on the US drinking-water Contaminant Candidate List 5.

Overview#

The chapter holds up well on its central argument and poorly on some specifics. Twenty-five years on:

This contradicts the chapter’s hedged statement (p. 117) that the groundwater threat “was never considered”. It was considered and not acted on. That makes the case a stronger example of discounted warnings than the chapter claimed. - The substitution chain continued, as the chapter warned (p. 118). - ETBE, the main European replacement, was also classified as IARC Group 2B in 2025. It has a lower taste-and-odour threshold in water than MTBE. - Corn-ethanol life-cycle greenhouse-gas savings remain contested, with some peer-reviewed estimates showing none. - The EU capped crop-based biofuels because of indirect land-use change.

2013 report. - The 2013 volume (Late lessons from early warnings: science, precaution, innovation) does not update this case. The annex of 2001-case summaries (Annex 2, report p. 742, in the EEA 2013 PDF I consulted online; the project brief calls it Annex 3) reproduces the 2001 summary and timeline unchanged. Its timeline still ends “2001 The debate goes on”. - Table A2.1 (p. 732) records MTBE’s “first early warning” as 1960 (taste, odour and persistence in water) and the action as “2000 undesirable in Denmark and California, permitted elsewhere”. It gives “40+” years of inaction. - The 2013 volume reuses the case in two places, again without post-2001 developments: - the costs “were borne by society” (p. 644); - alternatives bring their own hazards (p. 708).

Claim 1: Groundwater persistence and “everlasting” risk#

Original claim (pp. 111–112, 117, 120). - MTBE is highly soluble and mobile. - Field studies (Borden et al. 1997; Schirmer and Barker 1998) found no biodegradation. Significant degradation rates had “only been observed under laboratory conditions” (pp. 111–112). - Persistence “implies irreversibility” (p. 117). MTBE poses “an ongoing, everlasting risk of having irreversible adverse effects” (p. 120). - The chapter itself flags unpublished results suggesting biological degradation in ordinary waterworks sand filters (Arvin 2001, p. 114).

Subsequent developments. - Field biodegradation was demonstrated. - Kolhatkar et al. (2002, Environmental Science & Technology 36(23): 5139) used compound-specific carbon isotope analysis to show anaerobic MTBE biodegradation in groundwater at a gasoline release site. - US EPA’s research laboratory published a review of monitored natural attenuation of MTBE as a risk-management option at leaking tank sites (Wilson, Kaiser and Adair, January 2005, EPA/600/R-04/179). - An industry-sponsored technical protocol for evaluating MTBE natural attenuation followed in 2007 (API Publication 4761, May 2007). - Concawe’s 2012 European review (industry-funded) concluded that “it is now generally accepted that, after an initial acclimation period, GEO will biodegrade in the subsurface environment, albeit often at a lower rate than the BTEX compounds”. It also found that monitored natural attenuation is “a potentially viable risk management strategy” where there are no nearby receptors (Concawe report 4/12, 2012, Summary and §6.3). - A later review cites a survey of 50 European gasoline-impacted sites. Median MTBE and ETBE plume lengths (37 m and 38 m) were broadly comparable to those of the accompanying BTEX compounds (review in Journal of Hazardous Materials, 2020). - Biological treatment of drinking water became feasible at full scale. - Fluidised-bed bioreactors treated MTBE-contaminated supply water at full scale in Healdsburg and Laguna Hills, California. A pilot unit at Glennville kept effluent below detection limits (Schmidt et al., Journal of Hazardous Materials, 2012). - A field-scale fixed-bed bioreactor in North Hollywood, colonised by native aquifer bacteria, reported more than 99% removal (Biodegradation, 2013; abstract as indexed). - I could not locate a published version of the Danish waterworks sand-filter results cited as Arvin (2001). - Concentrations fell after the phase-out, but slowly in places. - USGS resampled north-eastern US well networks at ten-year intervals between 1996 and 2012 (Lindsey et al., Science of the Total Environment 579: 579–587, 2017). MTBE concentrations were “decreasing or unchanged in most wells as of 2012”, but still rising in a small number. The authors stress that these aquifers “show high susceptibility for contamination by a highly soluble, persistent constituent”. - In California, a public-water-supply database analysis (ES&T Letters 2(1): 7–11, 2015; authors at GSI Environmental) found: - MTBE had never been detected in over 98% of 13,183 tested public supply wells; - first-time detections peaked in 2000 and had fallen 80% since; - 142 of the 188 wells with any detection were non-detect at their latest test. - The USGS California decadal-trend report (Levy and Soldavini, 30 July 2026, SIR 2026-5039) found a statistically significant statewide decrease in MTBE over 2004–2023. It attributes the decrease to “decreased anthropogenic source inputs over time and natural attenuation in aquifers”. - Regulators still class MTBE as very persistent. - Anses’s 2023 regulatory management options analysis concluded that MTBE is not readily or inherently biodegradable. Its evidence was a soil DT50 above 180 days at 12 °C, citing the same Borden et al. (1997) study the chapter used, and groundwater DT50s of 50–200 days. Anses judged MTBE “very persistent” (vP) and, with a Koc of about 9 L/kg, “very mobile” (vM). - Anses also noted that treatment processes are “complex … and probably costly” (Anses opinion 2021-REACh-0093, 25 May 2023, §3.4).

Verdict: partly held up. - The core claim is confirmed by later regulatory assessment: high mobility plus slow degradation make MTBE a durable threat to groundwater resources. - The claim that significant degradation had been seen only in the laboratory was soon overtaken. Field biodegradation, natural attenuation and full-scale biological treatment are now established. - “Everlasting” was rhetorical overreach. The observed timescale is years to decades at contaminated sites, with regional concentrations declining after the source was removed.

Weight for the lessons. - The mechanism is sound: property combinations interacting with infrastructure (digest insight 4) and loss of a shared resource without toxicity (insight 5). Hindsight supports the idea that persistence and mobility together raise the cost of error. - The irreversibility argument (insight 11) should be weighted as “slow and costly to reverse”, not “irreversible”. Contamination is reversible in principle but costly and slow in practice. For example, Santa Monica’s main wellfield was offline from 1996 until at least the late 2000s. - Treat the chapter’s “only in the laboratory” framing as a snapshot of 2001 knowledge. It had in fact been partly contradicted by the chapter’s own citation of field evidence of anaerobic degradation (p. 111).

Claim 2: Carcinogenicity is borderline and unresolved#

Original claim (pp. 112–113, Box 11.1). - There was no pre-market cancer testing. - Three rodent bioassays showed tumours at multiple sites, but their relevance to humans was disputed. - Bodies reached different verdicts on the same data: - IARC (1999): Group 3, “not classifiable”; - US National Toxicology Program board: voted 6–5 against listing; - US EPA and the White House science council (NSTC): “carcinogenic potential”; - EU rapporteur: “borderline”; - European Chemicals Bureau working group (November 2000): rejected Category 3/R40.

Subsequent developments. - IARC reclassified MTBE upward. - An IARC working group met in February–March 2025 to evaluate automotive gasoline and oxygenated additives. It classified MTBE as Group 2B, “possibly carcinogenic to humans”, on “sufficient” evidence in experimental animals and “limited” mechanistic evidence. Human evidence was “inadequate” because no studies were available. - The animal tumours cited were: - hepatocellular adenoma or carcinoma in female CD-1 mice; - renal tubular adenoma or carcinoma in male F344 rats; - brain astrocytoma in male Wistar rats. - ETBE was also classified Group 2B, with “strong” mechanistic evidence. Tert-butyl alcohol (TBA), di-isopropyl ether (DIPE) and tert-amyl methyl ether (TAME) were placed in Group 3. - The summary appeared in The Lancet Oncology on 21 March 2025 (Turner et al.). IARC reports the full Volume 138 online as of 23 July 2026. - The same study was read in opposite ways. - The male Wistar astrocytomas come from a two-year drinking-water study (Dodd et al., Journal of Applied Toxicology, 2013). Incidences were 1/50, 1/50, 1/50 and 4/50 across dose groups. - The study’s authors concluded the tumours were “not associated with exposure to MTBE” (historical control range, no pairwise significance). - The ATSDR 2022 draft profile likewise describes the study as showing “no exposure-related tumors”. - IARC’s 2025 working group counted the finding towards “sufficient” animal evidence. - The EU has not classified MTBE as a carcinogen. - Its harmonised classification remains Flam. Liq. 2 (H225) and Skin Irrit. 2 (H315) (Anses opinions of 17 December 2020 and 25 May 2023). - Anses’s REACH substance evaluation (2014–2020) led to an ECHA decision (7 February 2017) requiring a transgenic rodent mutation assay (OECD TG 488). The study came back negative. Anses criticised its dose selection and the absence of confirmed tissue exposure, but accepted it as guideline-compliant. - Anses concluded that MTBE meets neither the CLP criteria for mutagenicity nor those for carcinogenicity. Its 2023 analysis added that this interpretation “could be reconsidered” in comparison with structurally similar substances. - The US has not classified MTBE. - EPA stated in 2008 that its MTBE health assessment was being revised (EPA 815-R-08-012, June 2008). - ATSDR’s 2022 draft reports that EPA (IRIS 1993) and HHS (NTP 2016) “have not classified” MTBE’s carcinogenic potential. - EPA did finalise IRIS assessments for ETBE and TBA in August 2021, each finding “suggestive evidence of carcinogenic potential” (as characterised in Rivera et al. 2025). - Industry-funded reviews argue against concern. - A 2025 systematic review (Borghoff et al., Current Research in Toxicology, February 2025) argues “a lack of concern for carcinogenic hazard in humans”. It was funded by LyondellBasell, Sustainable Fuels (formerly the European Fuel Oxygenates Association), the Asian Clean Fuels Association and a Latin American fuels association. One author is a LyondellBasell employee. - A companion review makes the same argument for ETBE and TBA (Rivera et al., November 2025). - WHO set no health-based guideline value. Its 2005 background document gives the reason: any such value would be far above the odour threshold of about 15 µg/l (WHO/SDE/WSH/05.08/122).

Verdict: held up. - The chapter’s description of an unresolved classification controversy, with divergent institutional verdicts on shared data, remains accurate a quarter-century later. - The direction of travel at IARC has been towards greater concern. EU and US regulatory classifications have not moved.

Weight for the lessons. - This strongly supports digest insights 9 (borderline evidence is decided by institutional defaults) and 10 (testing after deployment leaves uncertainty that outlives the decision to scale). Both can be weighted as strong. - The mechanism is now even more visible: - IARC’s hazard-identification criteria, EU CLP criteria (which weigh human relevance of modes of action) and industry weight-of-evidence reviews reach different answers from the same bioassays; - funding sources track conclusions. - Hindsight does not show that MTBE is a human carcinogen. The lesson is about how uncertainty is institutionally resolved, not about the underlying hazard.

Claim 3: “Near ignorance” on asthma and endocrine effects#

Original claim (pp. 112–113, 120). - A hypothesised link between MTBE combustion products and urban asthma rested on an editorial (Joseph 2000), conference papers, a newspaper report and local surveys. - A few rodent studies suggested endocrine effects, with no clear mechanism. - The chapter says a “state of near ignorance” characterises knowledge of both (p. 120).

Subsequent developments. - Asthma: not substantiated, and effectively dropped. - ATSDR’s toxicological profile (draft January 2022; final September 2023) says some occupational and population-based studies of the early-1990s oxygenated-fuel programmes observed respiratory symptoms and others did not, and “no clear conclusions can be drawn”. - Controlled human exposures at low levels showed no respiratory symptoms. - In ecological studies, asthma treatment was not increased in Anchorage or Fairbanks during the 1992–93 oxyfuel winter (Gordian et al. 1995). - A diagnosis of asthma was not increased in Philadelphia in 1997 compared with 1992 (Joseph and Weiner 2002), although a diagnosis of wheezing was. The 2002 study is by the same author whose 2000 editorial the chapter relied on. - Endocrine: still unresolved, with a new angle. - ATSDR found no human endocrine data. The adrenal gland may be a target at high doses. Male reproductive-hormone effects appeared only at doses causing overt toxicity and were “inconsistent across studies”. - Anses’s 2020 evaluation found “no mode of action … that emerges and reaches consensus”. - Its 2023 analysis concluded: - evidence for oestrogenic, androgenic, thyroid and steroidogenic (EATS) endocrine disruption was insufficient; - the environmental ED definition was not met (vitellogenin rose in male fish, but with no apical effects). - Through expert elicitation, however, Anses judged MTBE a “suspected endocrine disruptor for human health” through a metabolic pathway: insulin resistance via FABP4/PPARγ in adipocytes. This rested chiefly on one mouse study using a high-fat diet (Tang et al. 2019). - New harm categories have been raised. ATSDR notes a single cohort study reporting a possible association between MTBE exposure in the birth year and autism spectrum disorder (Kalkbrenner et al. 2018).

Verdict: partly held up. - “Near ignorance” was a fair description in 2001, and on endocrine effects it remains broadly fair. - The asthma hypothesis the chapter gave prominent space to (including the “alarmingly high” rates, p. 112) was not borne out. In retrospect the chapter over-weighted weak evidence. - The endocrine question was not closed. It shifted to a mechanism nobody was looking at in 2001, which supports the chapter’s call to reopen investigations “as new categories of adverse effects are discovered” (p. 120).

Weight for the lessons. - Treat the asthma material as an example of advocacy outrunning evidence. Do not use it as support for any lesson. - The endocrine trajectory supports a narrower lesson: hazard assessments of persistent, high-volume substances need periodic reopening as the toxicological toolkit changes. It does not show that such reopening will usually find harm.

Claim 4: US prediction of a shift to ethanol and non-oxygenated petrol#

Original claim (pp. 114–115). - California Executive Order D-5-99 (March 1999) called for removing MTBE “no later than December 31, 2002”. - US EPA announced in March 2000 that it would “significantly reduce or eliminate” MTBE. - The US was expected to shift to ethanol and to non-oxygenated petrol.

Subsequent developments. - California slipped a year. Governor Davis’s Executive Order D-52-02 (14 March 2002) postponed the ban by one year, citing supply and price risks. The Air Resources Board moved the deadline to 31 December 2003 (CARB; CRS RL32787). - Other states followed. New York banned MTBE from 1 January 2004. By 2006, 25 states had enacted full or partial restrictions, with dates ranging from 2000 to 2009 (CRS RL32787, updated 14 April 2006). - EPA’s route failed. EPA’s March 2000 initiative to restrict MTBE was never finalised. CRS reports that EPA had quietly dropped the effort by 2005. - Congress provided the federal exit. The Energy Policy Act of 2005 (P.L. 109-58, 8 August 2005): - §1504 struck the reformulated-gasoline oxygen-content requirement, effective 6 May 2006 per CRS; - enacted no federal MTBE ban; - dropped, in conference, a House “safe harbor” from product-liability suits; - created a Renewable Fuel Standard (4.0 billion gallons in 2006, rising to 7.5 billion by 2012). - The market exited quickly. - Valero announced on 2 August 2005 that it would stop producing MTBE in May 2006 (CRS). - MTBE’s share of reformulated-gasoline oxygenate fell from 87% (1999) to 46% (2004) (CRS). - USGS records that MTBE use “peaked in 1999 and was largely discontinued by 2007” (Lindsey et al. 2017). - IARC (2025) notes MTBE is “no longer added to gasoline in the USA” but is used “in Europe, Asia, and elsewhere”.

Verdict: held up. The direction and approximate timing were right. The mechanism differed: - state bans rather than a federal ban; - repeal of the oxygen mandate the Blue Ribbon Panel had recommended in 1999, achieved six years later; - liability exposure after the safe harbor failed; - a new mandate that locked in ethanol.

The shift to “non-oxygenated petrol” was largely pre-empted by the Renewable Fuel Standard.

Weight for the lessons. - This supports the chapter’s reading of a substitution chain (lead, then MTBE, then ethanol). - The exit shows that removing a mandated component took about seven years from the first state executive order and required legislation. That is useful evidence for the lock-in and mandate dynamics in digest insights 6 and 13.

Claim 5: EU prediction that MTBE keeps its place at low levels and tank rules keep risk low#

Original claim (p. 115). - EU reports anticipated that MTBE octane use would “settle out in the 1–4 percent volume range”. - Underground-tank rules were expected to take full effect by 2005 and keep groundwater risk “low in the future”, conditional on “strong enforcement”. - Benzene and aromatics cuts under Directive 98/70 “could result in an increase in the use of MTBE”. - The chapter characterises the EU position as treating leakage as “a technical problem” (p. 115).

Subsequent developments. - Risk reduction was left to Member States. Commission Recommendation 2001/838/EC (7 November 2001) adopted risk-reduction strategies for MTBE for Member States to implement. According to the indexed text, it recommends groundwater “monitoring programmes … where appropriate, in order to permit the early detection of groundwater contaminated by MTBE”. I could not open the full text (EUR-Lex blocked automated access). - The EU never restricted MTBE. Directive 2009/30/EC set a maximum of 22% v/v for ethers in petrol (Anses 2023, §3.2). - Actual levels were higher than anticipated. In 1,239 SGS sampling events across EU countries in 2000–09, mean MTBE content was 5.39% m/m and the median 4.25%, with a maximum of 20.4%. MTBE was detected in premium unleaded petrol in every EU-27 country in 2004–09 (Concawe 4/12, Table 2.1 and §2.5.1). - ETBE displaced MTBE for biofuel reasons. - ETBE’s share of European ether production capacity rose from 15% (2002) to about 60% (2010). Total ether capacity grew from 4.1 to 6.0 Mt/yr. - The driver was biofuel policy: ETBE made from bio-ethanol counts as a biofuel (Concawe 4/12, §§2.2, 2.6). - France switched from MTBE to ETBE from 2005, and Denmark was the first country to phase out MTBE (Anses 2023; Concawe 4/12). - MTBE remains registered under REACH in the 1–10 Mt/yr band. Anses reported at least one French plant producing more than 600,000 t/yr (Anses 2023, §3.3). - Groundwater monitoring stayed patchy. - Concawe’s industry-funded 2012 survey obtained data from only 11 of 30 European countries; 19 had “no specific data”. - Where data existed, background groundwater was “typically non-detectable”, impacts were localised point-source releases, and drinking water carried traces of 0.01–0.2 µg/l, with nothing above 20 µg/l reported. - Neither the Groundwater Directive (2006/118/EC) nor the priority-substances directive (2008/105/EC) sets a value for any fuel ether. Only Denmark had a legally binding drinking-water standard for MTBE, 5 µg/l (2007). RIVM proposed 1 µg/l for the Netherlands in 2004 (Concawe 4/12, §1.6). - Anses (2023) found that: - MTBE was detected in 20.7% of groundwater measurements in the NORMAN database; - MTBE was not on the EU Water Framework Directive watch list; - it “strongly recommends” EU-wide monitoring. - Anses also found potentially unacceptable local aquatic risks from fuel distribution and storage, and from industrial and professional uses (§§3.6, 4). - Tank-rule enforcement could not be assessed. I found no EU-level evaluation of tank enforcement. By contrast, the US legislated mandatory inspections of every regulated tank at least every three years, operator training and secondary containment in the Underground Storage Tank Compliance Act of 2005 (EPAct 2005, Title XV Subtitle B; US EPA). This suggests the enforcement concern flagged on p. 115 was well founded.

Verdict: partly held up. - MTBE did keep its place in the EU. - No US-scale drinking-water crisis has been documented in Europe, though the monitoring base is too thin to be confident. - The 1–4% range was exceeded on average in the following decade. - Substitution came from climate and biofuel policy rather than groundwater risk management. - By 2023 the French evaluating authority had concluded that MTBE’s properties warrant formal hazard identification and EU monitoring. That is closer to the chapter’s precautionary stance than to the 2001 “technical problem” framing.

Weight for the lessons. - Europe’s lower MTBE concentrations, lower groundwater reliance in some countries and tank rules appear to have limited damage. That tempers any claim that the US experience was inevitable. - The absence of systematic monitoring means “no documented problem” cannot be read as “no problem”. This supports digest insight 8 (missing monitoring masks problems). - On the evidence available, insight 12 (containment depends on enforcement) stays moderate.

Claim 6: California cost-benefit analysis showing MTBE as the most expensive option#

Original claim (pp. 114, 118–119). A 1998 University of California analysis (Keller et al.) found: - air-quality benefits of USD 14–78m/yr, similar across options; - MTBE water-treatment costs of USD 340–1,480m/yr in California, assuming granular activated carbon treatment to below 5 µg/l; - non-oxygenated petrol cheapest, then ethanol.

The chapter notes that cheaper biological treatment might lower MTBE’s costs (pp. 114, 119).

Subsequent developments. - The decision the analysis supported was taken. California dropped MTBE at the end of 2003, so the counterfactual of continued MTBE use was never observed. - Public-supply impacts in California proved narrower than the high estimates imply. - More than 98% of 13,183 public supply wells never detected MTBE, and first-time detections peaked in 2000 (ES&T Letters, 2015). - Over 2000–2009, 69 drinking-water sources reported detections above 3 µg/l and 24 exceeded the 13 µg/l primary limit (California State Water Resources Control Board). - Damage was concentrated and expensive where it occurred. - Santa Monica’s Charnock wellfield, about half the city’s supply, was shut in 1996. - The city recovered about USD 250m from Shell, ChevronTexaco and ExxonMobil (Heal the Bay; Santa Monica Mirror, March 2008). - US EPA reported that remediation spending was expected to exceed USD 200m, and that replacement water cost more than USD 3m a year (US EPA news release, 16 February 2005). - National cost estimates spanned two orders of magnitude. The water industry’s estimates were: - USD 25bn (American Water Works Association); - USD 33.2bn (Association of Metropolitan Water Agencies).

The petroleum industry’s estimate was USD 0.5–1.5bn (API). About 60% of gasoline-contaminated sites surveyed in 2003 contained MTBE (CRS RL32787, 2006). - Biological treatment became available. It proved workable at full scale (see Claim 1). That bears out the chapter’s caveat about cheaper treatment.

Verdict: partly held up. - The qualitative ranking held. The water-treatment costs attributable to MTBE dwarfed its component-specific air benefits, and policy followed. - The upper end of the Californian treatment-cost range now looks high relative to the public-supply impacts actually observed. Part of the reason is that the phase-out removed the source, so the forecast was partly self-negating. - The very wide range between water-industry and petroleum-industry national estimates shows how contested the cost accounting was.

Weight for the lessons. - This supports digest insight 7: component-specific benefits should be tested before a component is mandated. The asymmetry between small, shrinking benefits and large, persistent costs held. - Cite the specific Californian cost figures as an ex-ante estimate under continued use, not as realised costs.

Claim 7: Alternatives have their own problems#

Original claim (p. 118). - Ethanol “may not lower greenhouse gas emissions significantly” over its life cycle, and carries land, fertiliser and pesticide impacts. - Other ethers “could have health effects similar to those of MTBE”. - The University of California panel’s concern about “incomplete combustion products of ethanol” is noted (p. 114).

Subsequent developments. - ETBE shares MTBE’s classification and some of its water problems. - IARC classified ETBE Group 2B in 2025, the same as MTBE but with “strong” mechanistic evidence. - US EPA’s final IRIS assessment (August 2021) found “suggestive evidence of carcinogenic potential” for ETBE (as characterised by Rivera et al., Current Research in Toxicology, November 2025, an industry-funded review arguing lack of concern). - ETBE’s odour and flavour thresholds in water (1–2 µg/l) are lower than MTBE’s. It has been found near those thresholds in Dutch surface-water abstraction areas (van Wezel et al., Chemosphere 76, 2009). - Fate studies note that ETBE plumes are usually smaller than MTBE’s, largely because ETBE has been used in smaller amounts and is less soluble. The literature summarised in the 2020 ETBE review also indicates that alternative ethers would pose comparable groundwater threats if used at MTBE-like scale. - EPAct 2005 §1505 ordered EPA to study the public-health, air-quality and water-resource effects of ETBE, TAME, DIPE, TBA, ethanol, iso-octane and alkylates as MTBE substitutes. I did not verify whether or when that study was delivered. - Ethanol’s greenhouse-gas benefit is contested. - US EPA’s 2010 Renewable Fuel Standard lifecycle analysis estimated about a 21% reduction relative to gasoline. - Lark et al. (PNAS 119(9), February 2022) estimated that, once domestic land-use change is counted, corn ethanol under the Standard is likely at least 24% more carbon-intensive than gasoline. They also reported fertiliser and water-quality effects. - Their method was contested in a December 2022 PNAS letter (Alarcon Falconi et al.), which argues that accepted models give lower land-use-change emissions. - EPA found net environmental costs. EPA’s Biofuels and the Environment: Third Triennial Report to Congress (final, January 2025) concluded that the Standard had “a modest positive effect on biofuel production and consumption and thus had a modest negative effect on the environment”, across air and water quality, water quantity, ecosystems, soil and international impacts (Federal Register, 21 January 2025). - The EU capped crop-based biofuels. - Directive (EU) 2015/1513 acknowledged that indirect land-use change can negate some or all of the greenhouse-gas savings of individual biofuels. It capped food- and feed-based biofuels at 7% of transport energy. - RED II (Directive (EU) 2018/2001) and Delegated Regulation (EU) 2019/807 phase high-ILUC-risk feedstocks down to zero by 2030.

Verdict: held up. - Both cautions were borne out. - The chapter’s ether warning was vindicated by the 2025 IARC classification of ETBE and by ETBE’s water-quality profile. - Its ethanol caution became a mainstream and still unresolved policy controversy, which led to EU legal caps.

Weight for the lessons. - This strengthens digest insight 1 (evaluate substitutes on their own terms, not only against the incumbent). It can be weighted strong rather than moderate–strong. - It also shows the chapter’s point on p. 118 in action: substitutes chosen quickly under a mandate tend to be close chemical relatives, or they carry a different, displaced burden.

Claim 8: Full multi-media assessment before introduction; investigate persistent chemicals; seek alternatives#

Original claim. - The University of California task force and EPA’s Blue Ribbon Panel called for a “full, multi-media assessment” of any major new fuel additive before it is introduced (p. 114). - The chapter recommends “systematic, comprehensive and thorough investigations into all known possible adverse effects … before any release of large volumes of a persistent chemical”, “reopened and pursued as new categories of adverse effects are discovered”. - Alternatives to persistent chemicals should be sought “whenever possible” (p. 120). - The chapter asks whether “persistency alone” justifies precaution (p. 120).

Subsequent developments. - EU: the grandfathering gap was closed. - REACH (Regulation (EC) 1907/2006) ended the new-versus-existing-substance distinction the chapter identifies on p. 116. Existing substances like MTBE had to be registered with data. - MTBE was then placed on the Community Rolling Action Plan for 2014, with France as evaluating state. The initial concerns were suspected endocrine disruption, high tonnage and dispersive use; mutagenicity, biodegradability and persistence were added during the evaluation (ECHA CoRAP list 2014–2016; Anses 2020). - This is close to the “reopened … as new categories of adverse effects are discovered” process the chapter recommends. It came about 14 years after the chapter and some 40 years after marketing began. - EU: persistence and mobility became hazards in their own right. - Commission Delegated Regulation (EU) 2023/707 (published 31 March 2023) created CLP hazard classes for PMT and vPvM substances. A vPvM classification requires no toxicity criterion. It applies to substances newly placed on the market from 1 May 2025 and to those already on the market from 1 November 2026. - The rationale, set out by the German Environment Agency, is protection of drinking-water sources from substances that are persistent and mobile. That is precisely the MTBE problem. - Anses (2023) concluded that MTBE meets the vPvM criteria. It judged identification as a substance of very high concern, or harmonised classification, to be appropriate. - Anses also noted a major gap: REACH authorisation does not apply to uses as fuel (Art. 56(4)), MTBE’s dominant use. It suggested a REACH restriction could follow if updated registrant data confirmed the modelled risks. - As of the last public information I could access, MTBE’s harmonised classification had not been amended. I could not verify whether a CLH proposal or SVHC dossier has since been filed (ECHA pages were inaccessible). - US: fuel-additive testing became mandatory, and substitutes got retrospective review. - EPAct 2005 §1505 amended Clean Air Act §211(b) so that EPA “shall, on a regular basis” require fuel and additive manufacturers “to conduct tests to determine potential public health and environmental effects … (including carcinogenic, teratogenic, or mutagenic effects)”. It also mandated a study of the air, water and health effects of MTBE substitutes. - The Energy Independence and Security Act 2007 required triennial reports to Congress on the environmental impacts of the Renewable Fuel Standard. The third report was issued in January 2025. - The multi-media assessment of ethanol, the main replacement, came after the statutory mandates of 2005 and 2007 had scaled it, not before.

Verdict: strengthened. - The chapter’s central normative question has received a significant institutional answer in EU law: persistence combined with mobility is now a hazard class independent of toxicity. - Grandfathering of existing substances, which the chapter identifies as the gap MTBE fell through, has been closed in principle by REACH. - Implementation remains partial. Fuel uses escape authorisation, MTBE itself had not been formally classified vPvM by 2023, and the US repeated the pattern of scaling a replacement by mandate before full multi-media assessment.

Weight for the lessons. - Digest insight 2 (regimes that cover only new entrants leave established items unexamined) is strengthened. REACH’s reopening of MTBE produced new concerns within a few years. - Insight 3 (deployment scale should trigger scrutiny) is strengthened. REACH evaluation was triggered by tonnage and dispersive use. - The US ethanol sequence is a counter-example. Mandated scale-up can still outrun assessment even after the lesson has been written into reports, which supports insight 6 as strong.

Claim 9: Costs borne by society; unclear who should pay#

Original claim (p. 119). - “Until now, the majority of the costs … have been borne by society.” - “Who is the polluter, who should pay?” Both industry and regulators were “short-sighted”.

Subsequent developments. - Congress refused liability protection for producers. A House “safe harbor” shielding MTBE producers from defective-product claims was dropped from EPAct 2005 in conference (CRS RL32873; RL32787). - The courts assigned substantial liability to producers. - Multi-district litigation. In 2008, major oil companies settled claims by 153 public water systems in California and 19 other states for USD 423m, plus treatment of newly contaminated wells for 30 years. The court was the Southern District of New York (In re MTBE Products Liability Litigation, MDL 1358); the settlement was approved in August 2008 (Weitz & Luxenberg; SEC filings of settling companies, 2008). - New York City. A jury found ExxonMobil liable in October 2009 for public nuisance, negligence, trespass and failure to warn, and awarded USD 104.69m. The award was affirmed by the Second Circuit in July 2013, and the Supreme Court denied certiorari in April 2014 (NYC Law Department; FindLaw). - New Hampshire. A jury in April 2013 found ExxonMobil liable on design-defect and failure-to-warn theories. The New Hampshire Supreme Court affirmed USD 236.4m on 2 October 2015. It adopted market-share liability because commingled petrol cannot be traced to a supplier, and rejected a Clean Air Act pre-emption defence (State v. Exxon Mobil Corp., 168 N.H. 211). The US Supreme Court subsequently declined review (NH Department of Justice). - Santa Monica. The city received about USD 250m from Shell, ChevronTexaco and ExxonMobil following a November 2003 settlement, renegotiated in 2006. Contingency-fee counsel took USD 55m, or 22%, of the recovery (Santa Monica Mirror, 11 March 2008). - New Jersey. DEP sued nearly 50 companies in 2007 and had obtained more than USD 350m in settlements by March 2018. ExxonMobil remained a defendant (Insider NJ, 12 March 2018). - Pennsylvania and Maryland. State suits filed in July 2014 and December 2017 against more than 50 and more than 60 defendants were still unresolved as of Getty Realty’s 10-K filed in February 2026. Some defendants had settled. The Pennsylvania plaintiffs include the state’s Underground Storage Tank Indemnification Fund. - Public money still bore part of the cost. EPAct 2005 authorised USD 200m from the federal Leaking Underground Storage Tank Trust Fund specifically for tank leaks involving MTBE or other oxygenates (CRS RL32873). - Congress’s own findings framed industry as responding to a mandate. EPAct 2005 §1502 records that industry “responded to the fuel oxygenate standard … by making substantial investments in MTBE production capacity”. That framing is consistent with the chapter’s view that responsibility was shared.

Verdict: partly held up. - In 2001 the statement was accurate. - Afterwards, US courts supplied a substantial, if partial and slow, answer to “who should pay”. Producers were held liable, often on failure-to-warn theories grounded in their own early knowledge (see Claim 10). - Recoveries run to at least the low billions of dollars in aggregate. They remain well below the water utilities’ national cost estimates, and litigation was still running 30 years after Santa Monica closed its wells. Much of the cost therefore still fell on public budgets and ratepayers. - Outside the US, I found no comparable liability outcome. - The 2013 report repeated the 2001 claim without these updates (p. 644).

Weight for the lessons. - Digest insight 13 (mandated activities diffuse liability onto the public) needs qualifying. - A regulatory mandate did not shield producers where they had chosen among compliant options and had internal knowledge of the risk. The New Hampshire court rejected pre-emption on exactly this point. - The lesson is better stated in two parts: - liability diffuses by default; - it can be partly reassigned after the fact where documentary evidence of early knowledge exists and courts accept collective-liability theories. The process takes decades, and its reach depends on the jurisdiction.

Claim 10: Regulators ignored the 1990 warnings until 1996, a delay “documented in the United States and Denmark”#

Original claim (pp. 114–115). - “The first warnings” were the 1990 laboratory-scale papers (Barker et al.; Jensen and Arvin). Regulators “did not react … until the full-scale results emerged in 1996”. - Separately, the chapter says “it seems that the possibility that MTBE could pose a threat to groundwater reservoirs was never considered” (p. 117). - The Danish EPA’s admission is documented (p. 114). No US documentation is cited.

Subsequent developments. - US EPA put the groundwater concern in the Federal Register in 1988. The Testing Consent Order under TSCA section 4 (53 FR 10391, 31 March 1988) responded to an Interagency Testing Committee recommendation of 14 November 1986, which concerned vapour exposure and the lack of chronic data. EPA went further and wrote:

“EPA has an additional concern about MTBE contamination of ground water … the rapid growth in production, transport, and use of MTBE will probably contribute to an increase in incidents of contamination.”

The notice also records: - MTBE’s high solubility; - that an estimated 35% of about 638,000 non-farm motor-fuel tanks would fail a tightness test; - contamination of Rockaway Township, New Jersey’s wells (population 20,000) at 25–40 ppb, requiring treatment; - household-well contamination in Maine of up to 690 ppb; - maximum contaminant levels of 50 ppb already adopted in Maine and New Jersey.

It cites Garrett, Moreau and Lowry’s 1986 NWWA/API conference paper “Methyl tertiary butyl ether as a ground water contaminant”. The consent order led to the industry-sponsored Bushy Run cancer bioassays of 1992, which the chapter itself cites (pp. 113, 122). - Industry knew earlier still. - The New Hampshire Supreme Court’s 2015 opinion summarises trial evidence of an August 1984 memo by Exxon environmental engineer Barbara Mickelson. The memo predicted that “the number of well contamination incidents is estimated to increase three times” with MTBE, with clean-up costs “higher by a factor of 5”. - CRS dates an Exxon engineer’s memo making the same predictions to August 1985 and mentions a separate 1984 oil-company estimate. The two sources differ on the year. - The court also records that EPA raised concerns at 1986 industry MTBE Committee meetings that MTBE “can find its way to drinking supplies (i.e., aquifers)”. - A Tosco executive testified that his company decided against MTBE because of contamination and liability risk. - The 1990 Clean Air Act oxygenate requirement was framed around air quality alone. CRS notes that the Act did not mandate MTBE specifically, since ethanol could also meet the oxygen requirement. It was MTBE’s price and handling characteristics that led to its widespread use (RL32787, 2006). I found no evidence that groundwater was weighed in the 1990 decision. - The 2013 report’s timeline is unchanged. It still gives 1990 as the first groundwater indication (Table A2.11, p. 742).

Verdict: strengthened. - The chapter’s claim of regulatory delay is borne out and the documented delay is longer. In the US, warnings from a state agency (1986), the federal regulator (1988) and inside industry (1984–85) came well before the 1990 papers and roughly a decade before the 1996 USGS report. - Two specific statements are overturned: - that the first warnings came in 1990; - the hedged claim that the groundwater threat “was never considered” (p. 117). - The record shows the threat was considered, written down and discounted. The chapter’s argument that competent chemists “could have anticipated” persistence was hedged: “Documentation for this argument has however not been found” (p. 115). Later documents supply that documentation, from the industry and from US EPA.

Weight for the lessons. - Digest insight 8 (small-scale warnings are discounted until harm is large) should be upgraded from moderate to strong for this case. - The mechanism needs restating. The MTBE story is less about a failure of foresight and more about knowledge that existed inside firms and agencies failing to change a decision that was already under way. At the time, the benefits of the decision were concentrated and immediate, and its costs diffuse and deferred. - Recovering the internal documents through litigation, not through regulatory transparency, is itself a lesson. Much of what “was known” became public only through discovery.

Cross-cutting implications for using this chapter as a lens#

  1. Stronger than the chapter claimed: - substitute evaluation (insight 1); - grandfathering (insight 2); - decisions on borderline evidence set by institutional defaults (insights 9–10); - discounting of warnings (insight 8).

Later evidence strengthens all of these, and the regulatory-delay story is better documented now than in 2001. 2. Weaker than the chapter claimed: - irreversibility (“everlasting”); - the asthma hypothesis; - the specific foresight chronology.

Use persistence as “slow and costly to reverse”, and treat the asthma material as a cautionary example of weak evidence given prominent space. 3. Changed since 2001: - liability (insight 13) was partly reassigned by courts; - persistence plus mobility has become a legal hazard class in the EU.

Any use of the chapter should reflect both developments rather than the 2001 snapshot. 4. Recurring pattern: - lead was replaced by MTBE, MTBE by ethanol and ETBE, and each replacement was scaled by mandate or market before full assessment; - the replacements brought their own contested harms (ETBE’s classification and water profile; ethanol’s land-use and greenhouse-gas debate).

This is the most durable transferable lesson and is well supported by post-2001 evidence.

Sources#

Chapter and 2013 follow-up - EEA (2001) Late lessons from early warnings: the precautionary principle 1896–2000, Ch. 11 (pp. 110–125). https://www.eea.europa.eu/en/analysis/publications/environmental_issue_report_2001_22/issue-22-part-11.pdf/@@download/file - EEA (2013) Late lessons from early warnings: science, precaution, innovation, EEA Report 1/2013. Pp. 575, 644, 657, 708, 732 (Table A2.1) and 742 (MTBE annex summary, Table A2.11). Web copy consulted: https://eclass.hua.gr/modules/document/file.php/GEO168/EEA,%20SCIENCE,%20PRECAUTION,%20INNOVATION,%202013.pdf. EEA landing page: https://www.eea.europa.eu/en/analysis/publications/late-lessons-2

Carcinogenicity, toxicology and health - IARC (21 March 2025) “IARC Monographs evaluation of the carcinogenicity of automotive gasoline and some oxygenated gasoline additives” (news release). https://www.iarc.who.int/news-events/iarc-monographs-evaluation-of-the-carcinogenicity-of-automotive-gasoline-and-some-oxygenated-gasoline-additives - Turner MC et al. (2025) “Carcinogenicity of automotive gasoline and some oxygenated gasoline additives”, Lancet Oncology (published 21 March 2025). https://doi.org/10.1016/S1470-2045(25)00165-2 ; PMC full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC12303572/ - IARC Monographs Volume 138 page (full volume announced online 23 July 2026). https://monographs.iarc.who.int/news-events/volume-138-automotive-gasoline-and-some-oxygenated-additives - IARC (1999) Vol. 73 summary: MTBE Group 3. https://inchem.org/documents/iarc/vol73/73-13.html - Dodd DE et al. (2013) “Two-year drinking water carcinogenicity study of MTBE in Wistar rats”, Journal of Applied Toxicology. https://analyticalsciencejournals.onlinelibrary.wiley.com/doi/10.1002/jat.1776 - ATSDR (September 2023) Toxicological Profile for Methyl tert-Butyl Ether (MTBE), final. https://www.atsdr.cdc.gov/ToxProfiles/tp91.pdf. The public-comment draft of January 2022, from which the passages here were read, is at https://www.govinfo.gov/content/pkg/GOVPUB-HE20_500-PURL-gpo177762/pdf/GOVPUB-HE20_500-PURL-gpo177762.pdf - Borghoff SJ et al. (February 2025) “Systematic evaluation of the evidence base on methyl tert-butyl ether supporting a lack of concern for carcinogenic hazard in humans…”, Current Research in Toxicology (industry-funded). https://www.sciencedirect.com/science/article/pii/S2666027X25000106 ; PubMed: https://pubmed.ncbi.nlm.nih.gov/40084233/ - Rivera BN et al. (10 November 2025) “Systematic evaluation of the evidence base on ETBE and TBA…”, Current Research in Toxicology 10: 100270 (industry-funded). https://pmc.ncbi.nlm.nih.gov/articles/PMC12723127/ - US EPA IRIS, ETBE final assessment (August 2021). https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=348436 ; TBA final assessment (August 2021): https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=348442 - WHO (2005) Methyl tertiary-Butyl Ether (MTBE) in Drinking-water, WHO/SDE/WSH/05.08/122. https://www.who.int/docs/default-source/wash-documents/wash-chemicals/methyl-tertiary-butyl-ether-background-document.pdf

EU chemicals regulation - Anses (17 December 2020) Avis … relatif à l’évaluation du MTBE dans le cadre du règlement REACH (saisine 2019-SA-0153). https://www.anses.fr/en/system/files/REACH2019SA0153.pdf - Anses (25 May 2023) Avis … relatif à l’analyse des options de gestion réglementaires du MTBE (saisine 2021-REACh-0093). https://www.anses.fr/fr/system/files/REACH2021REACh0093.pdf - ECHA CoRAP list 2014–2016 (MTBE: 2014, France). Mirror: https://www.crad.com.tr/UPLOAD/URUN/FILES/CoRAP2014_2016-0717402192.pdf ; ECHA CoRAP page (not accessible to me): https://echa.europa.eu/information-on-chemicals/evaluation/community-rolling-action-plan/corap-table - ECHA substance evaluation conclusion document for MTBE (not accessible to me; cited by Anses 2023). https://echa.europa.eu/documents/10162/dbafd544-5671-8b03-980c-37a2bc9291b0 - Commission Delegated Regulation (EU) 2023/707 (19 December 2022; published in the OJ 31 March 2023), new CLP hazard classes. https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32023R0707 - German Environment Agency (UBA), “PMT and vPvM substances under REACH”. https://www.umweltbundesamt.de/en/PMT-substances - Commission Recommendation 2001/838/EC of 7 November 2001 (MTBE risk-reduction strategy; indexed excerpt only). https://eur-lex.europa.eu/eli/reco/2001/838/oj/eng

European fuel use and groundwater occurrence - Concawe (2012) Gasoline ether oxygenate occurrence in Europe, and a review of their fate and transport characteristics in the environment, report no. 4/12 (industry). https://www.concawe.eu/wp-content/uploads/report-no-4_12.pdf - van Wezel A et al. (2009) “Odour and flavour thresholds of gasoline additives (MTBE, ETBE and TAME) and their occurrence in Dutch drinking water collection areas”, Chemosphere 76. https://pubmed.ncbi.nlm.nih.gov/19477481/ - Review (2020) “Biodegradation and fate of ethyl tert-butyl ether (ETBE) in soil and groundwater: A review”, Journal of Hazardous Materials. https://www.sciencedirect.com/science/article/pii/S0304389420300327

Groundwater trends, biodegradation and treatment - Lindsey BD, Ayotte JD, Jurgens BC, DeSimone LA (2017) Science of the Total Environment 579: 579–587. https://pubs.usgs.gov/publication/70179869 - Levy ZF, Soldavini AL (30 July 2026) Decadal trends in the quality of groundwater used for public drinking-water supply in California, 2004–23, USGS SIR 2026-5039. https://pubs.usgs.gov/publication/sir20265039/full - “Life Cycle of Methyl tert-Butyl Ether in California Public Water Supply Wells” (2015), ES&T Letters 2(1): 7–11. https://pubs.acs.org/doi/10.1021/ez500377t ; summary: https://www.gsienv.com/gsi-papers/life-cycle-of-methyl-tert-butyl-ether-in-california-public-water-supply-wells/ - California State Water Resources Control Board, “MTBE: Regulations and Drinking Water Monitoring Results”. https://www.waterboards.ca.gov/drinking_water/certlic/drinkingwater/MTBE.html - Kolhatkar R et al. (2002) “Use of Compound-Specific Stable Carbon Isotope Analyses To Demonstrate Anaerobic Biodegradation of MTBE in Groundwater at a Gasoline Release Site”, ES&T 36(23): 5139. https://pubs.acs.org/esthag/article-abstract/36/23/5139/3707357/ - Wilson JT, Kaiser PM, Adair C (January 2005) Monitored Natural Attenuation of MTBE as a Risk Management Option at Leaking Underground Storage Tank Sites, EPA/600/R-04/179. https://cfpub.epa.gov/si/si_public_record_report.cfm?dirEntryId=135426&Lab=NRMRL - API (May 2007) Technical Protocol for Evaluating the Natural Attenuation of MtBE, Publication 4761 (industry). https://frtr.gov/matrix/documents/Monitored-Natural-Attenuation/2007-Technical-Protocol-for-Evaluating-the-Natural-Attenuation-of-MTBE.pdf - Schmidt R et al. (2012) “Microbial biosafety of pilot-scale bioreactor treating MTBE and TBA-contaminated drinking water supply”, Journal of Hazardous Materials. https://pmc.ncbi.nlm.nih.gov/articles/PMC3327512/ - “Successful treatment of an MTBE-impacted aquifer using a bioreactor self-colonized by native aquifer bacteria” (2013), Biodegradation (abstract as indexed). https://link.springer.com/article/10.1007/s10532-013-9639-0

US policy and regulation - US EPA (31 March 1988) “Testing Consent Order on Methyl Tert-Butyl Ether and Response to the Interagency Testing Committee”, 53 FR 10391. https://www.epa.gov/sites/default/files/2015-10/documents/sun91.pdf - McCarthy JE, Tiemann M, CRS RL32787 MTBE in Gasoline: Clean Air and Drinking Water Issues (updated 14 April 2006). https://www.everycrsreport.com/reports/RL32787.html - CRS RL32873 Key Environmental Issues in the Energy Policy Act of 2005. https://www.everycrsreport.com/reports/RL32873.html - Energy Policy Act of 2005, P.L. 109-58 (8 August 2005), §§1502–1509. https://www.govinfo.gov/content/pkg/PLAW-109publ58/pdf/PLAW-109publ58.pdf - US EPA, “Energy Policy Act of 2005 and Underground Storage Tanks”. https://www.epa.gov/ust/energy-policy-act-2005-and-underground-storage-tanks-usts - Governor Gray Davis, Executive Order D-52-02 (14 March 2002). https://ww2.arb.ca.gov/sites/default/files/classic/isd/fuels/gasoline/oxy/eo5202.pdf - US EPA archive, “MTBE: Drinking Water” (1997 advisory, 20–40 µg/l). https://archive.epa.gov/mtbe/web/html/water.html - US EPA (June 2008) Regulatory Determinations Support Document for CCL 2, Ch. 13 MTBE, EPA 815-R-08-012. https://www.epa.gov/sites/default/files/2014-09/documents/chapter_13_mtbe.pdf - US EPA (January 2025) Biofuels and the Environment: Third Triennial Report to Congress (final). https://assessments.epa.gov/biofuels/document/&deid=363940 ; Federal Register notice, 21 January 2025: https://www.federalregister.gov/documents/2025/01/21/2025-01385/notice-of-the-final-biofuels-and-the-environment-third-triennial-report-to-congress

Litigation and liability - State of New Hampshire v. Exxon Mobil Corp., N.H. Supreme Court, 2 October 2015. https://caselaw.findlaw.com/court/nh-supreme-court/1714764.html ; official PDF: https://www.courts.nh.gov/sites/g/files/ehbemt471/files/documents/2021-08/2015080exxon.pdf - NH Department of Justice, “U.S. Supreme Court Upholds $236 Million Verdict Against ExxonMobil for MTBE Contamination”. https://www.doj.nh.gov/news-and-media/us-supreme-court-upholds-236-million-verdict-against-exxonmobil-mtbe-contamination - In re MTBE Products Liability Litigation (2d Cir. 2013), City of New York v. ExxonMobil. https://caselaw.findlaw.com/court/us-2nd-circuit/1640078.html ; NYC Law Department release: https://www.nyc.gov/html/law/downloads/pdf/Exxon%20MTBE%20Appeals%20Win.pdf ; AMWA note on certiorari denial: https://www.amwa.net/article/supreme-court-declines-review-mtbe-judgment - $423m MDL settlement (2008): Weitz & Luxenberg (plaintiffs’ firm, secondary). https://www.weitzlux.com/firm-news/423-million-mtbe-settlement/ ; e.g., American Water Works Co. Form 10-Q FY2008: https://www.sec.gov/Archives/edgar/data/0001410636/000119312508227647/d10q.htm - Getty Realty Corp. Form 10-K for FY2025 (filed February 2026), Pennsylvania and Maryland MTBE litigation status. https://www.sec.gov/Archives/edgar/data/1052752/000119312526048965/gty-20251231.htm - Insider NJ (12 March 2018) “Grewal: $196 Million in MTBE Settlements with Sunoco, BP and Shell”. https://www.insidernj.com/grewal-196-million-mtbe-settlements-sunoco-bp-shell/ - US EPA news release (16 February 2005) “Oil companies pay U.S. EPA to settle Santa Monica MTBE cleanup costs”. https://www.epa.gov/archive/epapages/newsroom_archive/newsreleases/5a790d8bf1a74361852570d8005e16fc.html - Santa Monica Mirror (March 2008) “MtBE Settlement Agreement Reached with Attorneys”. https://smmirror.com/2008/03/mtbe-settlement-agreement-reached-with-attorneys ; Heal the Bay, “Well, a Big Win for Santa Monica”: https://healthebay.org/well-a-big-win-for-santa-monica/

Ethanol and biofuels - Lark TJ et al. (February 2022) “Environmental outcomes of the US Renewable Fuel Standard”, PNAS 119(9): e2101084119. https://www.pnas.org/doi/10.1073/pnas.2101084119 - Alarcon Falconi TM et al. (13 December 2022) “Inconsistencies in domestic land use change study”, PNAS (letter). https://pmc.ncbi.nlm.nih.gov/articles/PMC9907145/ - Directive (EU) 2015/1513 (ILUC Directive). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:32015L1513 ; Delegated Regulation (EU) 2019/807 (high-ILUC-risk feedstocks): https://eur-lex.europa.eu/eli/reg_del/2019/807/oj/eng