Late Lessons, Jensen Huang and AI
Markdown version

LL1-02 hindsight check: Ch. 2 “Fisheries: taking stock” (MacGarvin), Late lessons from early warnings (EEA 2001), pp. 17–30#

Checked 25 September 2026. The check covers evidence from 2001 to September 2026.

Method note. The first pass was done without general web search. Sources were retrieved directly: agency documents (DFO science advisory reports and fishery decisions, ICES advice and its assessment database, NOAA, EEA, EU legislation, the UN Fish Stocks Agreement), DOI-resolved abstracts of peer-reviewed papers, and official statistics. A second pass on the same day, with web search, added or corrected six things: - the 2013 EEA report’s own treatment of this chapter (its Annexes 2 and 3, and Chapter 17); - the full text of the 1990 Harris report; - the EU–Norway–UK agreed records for North Sea cod catch limits in 2024–26; - the ICES September 2025 advice sheet itself; - a correction to the North Sea cod catch-limit table under Claim 4; - the recent history of Barents Sea (Northeast Arctic) cod.

Secondary sources are used only where noted.

Three items could not be verified: - the chapter’s 0.1% return-on-capital figure (from Vol. 2 of the 2001 Green Paper); - the North Sea industrial-catch figure (2.2 of 3.5 Mt); - the 1939 replacement of California’s state experts. This rests on McEvoy (1986), whose full text could not be retrieved.

Numbers marked “(my arithmetic)” are derived from figures in the cited documents.

Overview#

Twenty-five years on, the chapter’s diagnostic claims have mostly held up, and several have been strengthened: - Stock assessments overestimate stock size and underestimate fishing mortality when stocks are declining. - Managers set catches above the scientific advice. - Precautionary language can sit on top of unchanged practice. - Stocks are structured into sub-populations that management ignores.

The North Sea is the strongest vindication. ICES advised “closure” or “zero catch” for North Sea cod in most years from 2003 to 2009, while total allowable catches (TACs) of roughly 20,000–29,000 t were set. The retrospective over-estimation the chapter flagged in 1999 was still visible in the September 2025 assessment. By 2025–26, ICES was advising zero catch for every cod stock from the North Sea to the Baltic. That includes the Southern North Sea substock, which has been below its biomass limit since 2017. For 2026, against that zero-catch advice, the EU, Norway and the UK set a whole-stock TAC of 14,034 t.

The chapter’s historical details are weaker in places: - The California sardine collapse is misdated. It came in the mid-to-late 1940s and early 1950s, not 1942. - Current science treats the sardine collapse as largely driven by the environment, with fishing making it worse. Whether a lower northern cod TAC in 1990 would have averted collapse remains unproven.

The chapter’s implied pessimism about model-based (“first-generation”) precaution is now contested: - Many assessed stocks worldwide have rebuilt under harvest rules set by models. The US has rebuilt 51 stocks since 2000, and on average assessed stocks are at target abundance. - Cod in changing ecosystems keeps failing under the same machinery. The main examples are the Gulf of Maine, Georges Bank, the Baltic and the southern North Sea.

The chapter’s proposed remedy of “second-generation” precaution (no-take zones and ecosystem-based management) has a mixed record: - The Georges Bank closures helped haddock and scallops, but not cod, and part of them was removed in 2018. - Ecosystem-based management is written into law almost everywhere but used in only a small fraction of real catch-setting.

Northern cod did not recover on the timescale the chapter implied. Commercial fishing reopened only in 2024, and the stock is now assessed as “Healthy”. The reopening rests heavily on a revised limit reference point. DFO’s own framework states that “The change in relative stock status results from a downward revision of the LRP, not an increase in the quantity of cod.” Its 2026 report adds that “perception of stock status has tended to improve with each assessment more than is supported by changes in absolute biomass alone.” This is a fresh instance of the chapter’s lesson about model dependence and how “precaution” gets framed. It is not simply a happy ending.

How the EEA’s 2013 report revisited this chapter#

The 2013 report (EEA Report 1/2013) did not update the chapter’s stock claims. There are three relevant pieces, all taken from the EEA’s own downloads of the report’s annexes and Chapter 17:

Implication. The EEA never re-examined the chapter’s evidence. Its 2013 treatment kept the sardine misdating and added a more favourable precautionary contrast case (Barents Sea cod). The later history of that case is mixed (see Claim 6).

Claim 1. Northern cod: a “deliberately conservative” regime failed because assessments overestimated the stock and underestimated fishing mortality; collapse, 1992 moratorium, costs “in excess of several billion Canadian dollars” (pp. 20–22)#

Original claim. After 1977 Canada aimed to take about 20% of the stock. Keats et al. (1986) put removals at 1.5–3 times that target. Alverson (1988) found fishing mortality “considerably in excess” of target because the stock size was consistently overestimated. Harris (1990) found fishing mortality more than double the intended level and the stock little more than half its assumed size. The collapse led to the July 1992 moratorium, and 1990s costs exceeded “several billion” CAD, a figure sourced to the author’s own WWF-UK report (MacGarvin 2001a).

Subsequent developments. - The Harris report itself confirms the chapter’s paraphrase. Its Executive Summary says new methods showed “the actual fishing mortality rates since 1977 had in fact been at least double those projected in the F0.1 strategy” (Harris 1990, Final Report, February 1990, p. 3: https://waves-vagues.dfo-mpo.gc.ca/library-bibliotheque/114276.pdf). The report also records that in late 1988 scientific advice was to halve the TAC from 266,000 t. The Minister instead chose “the temporary expedient” of 235,000 t (pp. 9–10). - Peer-reviewed reconstructions backed the overfishing diagnosis. - Hutchings and Myers (1994) concluded that “the collapse of northern cod can be attributed solely to overexploitation”. They rejected environmental explanations, noting that similar harvests were sustainable in colder periods (https://doi.org/10.1139/f94-214, 1994). - Myers, Hutchings and Barrowman (1997) found no drop in recruitment before the collapse. Across six collapsed stocks, high juvenile mortality went with high adult fishing mortality, which is consistent with unrecorded discarding and misreporting (https://doi.org/10.1890/1051-0761(1997)007[0091:WDFSCT]2.0.CO;2, 1997). - Shelton et al. (2006), a DFO-authored paper, called “excessive and unsustainable fishing mortality … the predominant factor” in the depletion of Northwest Atlantic cod (https://doi.org/10.1139/f05-253, 2006). - DFO’s current assessment model broadly confirms high historical fishing pressure but tells a different story about the early 1990s. - The model, extended back to 1954, finds that fishing mortality “exceeded [natural mortality] through most of the 1950s to the 1980s”. It also shows natural mortality spiking to about 2.5 in 1992–94 (DFO SAR 2025/043, Sept 2025: https://publications.gc.ca/collections/collection_2025/mpo-dfo/fs70-6/Fs70-6-2025-043-eng.pdf; DFO SAR 2026/030, July 2026: https://publications.gc.ca/collections/collection_2026/mpo-dfo/fs70-6/Fs70-6-2026-030-eng.pdf). - DFO’s 2023 framework review concedes that “a large portion of what is estimated to be M … could actually be F stemming from unreported landings (Rose and Walters 2019)”, or else starvation linked to the capelin collapse (DFO SAR 2024/046, Aug 2024: https://publications.gc.ca/collections/collection_2025/mpo-dfo/fs70-6/Fs70-6-2024-046-eng.pdf). - Rose and Walters (2019) re-ran the data with a different model. They concluded that fishing “was a major contributor to the early 1990s decline” and that the official model overestimated biomass by about 35% (https://doi.org/10.1016/j.fishres.2019.105314, 2019). - The scale of the costs is corroborated from other sources, though not the exact figure. - About 30,000 Newfoundland and Labrador fishers and plant workers lost their work, roughly 12% of the provincial labour force. It was “the single largest mass layoff in Canadian history”. - The second relief programme alone, TAGS, was a “$1.9-billion program” that ran out of money in May 1998. The earlier NCARP programme came on top of that (Heritage NL, Memorial University: https://www.heritage.nf.ca/articles/economy/moratorium-impacts.php, undated; this is a secondary source). - The same source notes that the province’s total landed value had recovered to $321M by 1995 (against $277M in 1990), because shellfish replaced cod. Communities and employment did not recover in the same way.

Verdict: held up. The central diagnosis (overestimated stock, underestimated fishing mortality, then collapse) is mainstream. The costs are credibly in the billions. The shape of the 1980s–90s decline is still disputed between models: DFO’s model makes it more abrupt and driven by natural mortality, while Rose and Walters make it more gradual and driven by fishing.

Weight for the lessons. The lesson about “hindsight-only estimates” and false reassurance during a decline can carry heavy weight. The lesson about the costs of inaction can carry heavy weight as to scale. Attribution needs care: even today, the relative share of fishing and environment depends on which model you use (see Claim 2).

Claim 2. Northern cod had not recovered by 2000; DFO admitted it did not know why it collapsed or failed to recover; the “stock” is made up of discrete local populations (p. 22)#

Original claim. Only a 9,000 t inshore fishery reopened in 1999. DFO’s 2000 assessment judged even that catch above the 20% reference level and admitted ignorance of the causes. There was growing recognition that the stock comprises discrete local populations.

Subsequent developments. - Status in 2000 and the 2003 closure. The directed inshore fishery was closed in 2003, and fishing mortality has stayed below 0.05 since 2004 (DFO SAR 2026/030). - A partial comeback from about 2008 to 2016. Rose and Rowe (2015) documented spawning biomass rising “from tens of thousands of tonnes to >200 thousand tonnes within the last decade”, tracking the recovery of capelin (https://doi.org/10.1139/cjfas-2015-0346, 2015). DFO’s model then shows spawning biomass flat from 2016 to 2024 (DFO SAR 2024/049, Sept 2024: https://publications.gc.ca/collections/collection_2025/mpo-dfo/fs70-6/Fs70-6-2024-049-eng.pdf; SAR 2026/030). - The 2023 reclassification rested on a new reference point. - The October 2023 framework replaced the old limit reference point (average 1980s spawning biomass) with 40% of the biomass at maximum sustainable yield (B_MSY), estimated from a model extended back to 1954. - Under the old point, 2021 biomass was 0.52 times the limit. Under the new one it was 1.16 times, with a 29% chance of being below it, although estimated biomass was about the same (411 kt under the old model, 368 kt under the new). - DFO states: “The change in relative stock status results from a downward revision of the LRP, not an increase in the quantity of cod estimated by the revised model” (SAR 2024/046). - Those figures imply the new limit is roughly 40% of the old one (my arithmetic). The 2024 assessment’s wording, “the previous LRP was roughly 40% higher than the current LRP” (SAR 2024/049), does not match them. This looks like a drafting slip. - The fishery reopened, and catches rose quickly.

Year Decision date TAC What the science said
2024 26 June 2024 (decision) 18,000 t (SAR 2025/043 gives 18,947 t; total authorised removals 21,317 t) SAR 2024/049: 62–76% risk of decline for removals of 0–27,033 t; “no level of removals” gave a ≥50% chance of growth
2025 18 June 2025 (decision) 38,000 t (total authorised removals 42,867 t, per SAR 2026/030) This removal level is essentially the top (2×) scenario in SAR 2025/043, which projected a 71% chance of decline and a 27% chance of reaching the Critical Zone by 2028
2026–27 12 June 2026 (decision) 59,000 t SAR 2026/030 puts 2026 biomass at 542 kt, 2.3 times the limit, in the “Healthy Zone” (70% chance of being above the new upper reference, 80% of B_MSY). Projections to 2029 give a 21–53% chance of decline across removals of 0–85,734 t

Verdict: held up. As a snapshot of 2000 the claim was accurate. The epistemic core is, if anything, strengthened: DFO still cannot fully say why the stock collapsed or why it recovered slowly, and it now documents model instability openly. “Failure to recover” should not be read as permanent, though. The stock partly rebuilt over about 30 years, and a commercial fishery has reopened.

Weight for the lessons. The northern cod story now supports the chapter’s model lock-in and framing lessons more strongly than in 2001. A stock’s official status can change mainly because its reference point changes. Managers have then taken catches at the upper end of the scenarios they were shown. The chapter’s lesson on humility and its positive note that “the assessments are freely available and the uncertainties clearly set out” (p. 22) are also borne out. DFO’s recent reports are unusually candid.

Original claim. Harris judged that cutting the TAC from 235,000 t (1989) to about 125,000 t would have “drastic” repercussions. He suggested 190,000 t while warning “this may contribute to further decline”, and similar limits followed for 1991–92. On a close reading, the chapter’s figures (CAD 26M of landings, CAD 66.6M of processed product, about 1,000 jobs) appear to describe the cost of the cut that was actually made, from 235,000 t to 190,000 t. They are not the larger cost Harris sought to avoid.

Subsequent developments. - What Harris actually said (primary source). The final report (February 1990) shows that the chapter’s summary is accurate but leaves out three things (Harris 1990: https://waves-vagues.dfo-mpo.gc.ca/library-bibliotheque/114276.pdf). - 190,000 t was a staging point, conditional on a further cut. The report warns that “a 1990 TAC of 190,000 metric tons … may not serve to reverse the trend of a declining spawning stock but may rather contribute to further decline”. It balances this against the “social and economic repercussions of a particularly drastic nature” of cutting F straight to 0.2. It then proposes F = 0.3 “as a staging point on the way to the lower figure”, and adds that if the autumn 1989 survey confirmed F of 0.4 or more, “a lower TAC for 1991 will be imperative”. Its closing warning: “the longer the delay in facing the brutal reality, the harder and longer will be the road back” (p. 136). - The Panel did not foresee collapse. It found “not an immediate threat to the survival of the northern cod stock”. It was “reasonably confident” that DFO’s new method put fishing mortality “in the right domain” (p. 3). This backs the chapter’s point that 1992 was “far worse than even the most pessimistic projections”. It also shows that the independent review shared the model confidence the chapter criticises. - The cost figures are not in the report’s text. The chapter’s CAD 26M, CAD 66.6M and 1,000 jobs could not be found in the report’s machine-readable text. They may come from elsewhere, or be lost in the scan. - What was actually set, and conflicting accounts. - A secondary account (Brubaker 2000: https://environment.probeinternational.org/2000/01/18/unnatural-disaster-how-politics-destroyed-canadas-atlantic-groundfisheries/, 18 Jan 2000) gives the 1990 TAC as 197,000 t, cut “by another 7,000 tonnes in 1991”. That is consistent with the chapter’s “similar limits were set for 1991–92”. - That means the “imperative” further cut for 1991 that Harris made conditional on the survey was not made on anything like the scale he implied. - The EEA’s own 2013 report says Harris’s recommended TAC reduction “was not approved” (McGlade and van den Hove 2013, p. 414). That conflicts with the 2001 chapter. Taken together, the figures suggest managers roughly adopted Harris’s staging level for 1990 and then did not follow through. - No study I retrieved tests the specific 125,000 t counterfactual. The evidence points both ways. - Against the idea that 125,000 t would have been enough: DFO’s current model attributes the steepest part of the 1991–94 collapse to a spike in natural mortality to about 2.5, linked partly to the 1991 capelin collapse and extreme cold (SAR 2024/046; SAR 2026/030). On that account, a lower 1990 TAC alone might not have prevented a severe decline. The chapter itself concedes the outcome exceeded every projection. - For the view that fishing restraint mattered: Rose and Walters (2019) conclude that fishing was “a major contributor to the early 1990s decline”. Hutchings and Myers (1994) attribute the collapse “solely to overexploitation”. Rose (2004) models the late-1980s collapse as caused by “both” fishing and climate (https://doi.org/10.1139/f04-173). Schijns et al. (2021) estimate that “if fishing effort and mortality had been stabilized in the 1980s, precautionary annual yields of about 200000 tonnes could have been sustained” (https://doi.org/10.1093/icesjms/fsab153). - The counterfactual that matters is earlier restraint, in the 1980s. Mohn (1999) showed that uncorrected retrospective bias in cod assessments “could lead to catch-level advice that would be twice or more the intended level” (https://doi.org/10.1006/jmsc.1999.0481). That is the same mechanism the chapter describes for 1977–89.

Verdict: partly held up. The account of the decision is consistent with the Harris report, which I have now checked directly. - The chapter leaves out that Harris made a further cut for 1991 “imperative” if fishing mortality stayed high. That cut was not made. This makes the failure less a scientist’s compromise and more a failure of management to follow through. - The implication that a lower 1990 TAC would have averted collapse remains unproven, and the chapter is itself candid about this. - The strongest later evidence points to restraint needed across the 1980s, not a single-year cut in 1990.

Weight for the lessons. The lesson on asymmetric proof and costs (weighing the certain, visible costs of cutting catches against the uncertain, larger costs of collapse) keeps moderate-to-strong weight as a description of how decisions were made. The DFO decisions of 2024–26 (Claim 2) show the pattern recurring. It should not be presented as proof that one decision caused the collapse.

Claim 4. North Sea cod, haddock and whiting showed the same underestimation of fishing mortality in a 1999 retrospective analysis, known since 1977; many stocks were below B_pa and cod near B_lim; the view that Northeast Atlantic cod can tolerate higher mortality is “circular” (pp. 24–25)#

Original claim. The chapter drew a parallel with northern cod and implied that North Sea cod faced a real risk of collapse.

Subsequent developments. - North Sea cod stayed below its biomass limit for about a decade. - ICES’s 2001 assessment put spawning biomass (SSB) at 53.7 kt in 2000 against a B_lim of 70 kt, with fishing mortality above 1.0 in 1999 (ICES assessment database, SAG assessmentKey 102: https://sag.ices.dk/SAG_API/api/SummaryTable?assessmentKey=102). - The 2011 assessment put the low at 29.4 kt in 2006 (key 112). - The 2019 assessment, with B_lim revised to 107 kt, still shows SSB below B_lim from about 1999 to 2013. It shows a partial recovery to about 120 kt in 2015–16 and a fall back to about 80 kt by 2019 (key 13211). - A warning of the collapse risk was already in the mainstream literature before the chapter: Cook, Sinclair and Stefánsson, “Potential collapse of North Sea cod stocks”, Nature 1997 (https://doi.org/10.1038/385521a0). Horwood et al. (2006) found that effort cuts had reduced fishing mortality by about 37% by 2005, “insufficient to ensure recovery … within the next decade” (https://doi.org/10.1016/j.icesjms.2006.05.001). - Managers repeatedly set TACs above closure advice. - The 2001–2009 rows below come from ICES’s own advice-history table, where the TAC is for the North Sea (Subarea 4) only (ICES Advice 2025, cod.27.46a7d20, 23 Sept 2025, Table 6a: https://doi.org/10.17895/ices.advice.27202566). - The 2024–2026 rows use whole-stock figures from the EU–Norway–UK agreed records. Since the 2023 benchmark, ICES advice covers the whole northern shelf stock, so this is the like-for-like comparison. - Correction to the first draft of this check. The first draft set ICES’s Subarea 4 figure of 19,910 t against whole-stock advice. That understated the gap. ICES’s table also lists 19,910 t for 2024, where the 2024 agreed record gives 24,900 t for Subarea 4.

Year ICES advice Agreed TAC
2001 “Lowest possible catch” 48,600 t (Subarea 4)
2003 “Closure” 27,300 t (Subarea 4)
2004–2007 “Zero catch” 27,300 t down to 19,957 t (Subarea 4)
2009 “Zero catch” 28,798 t (Subarea 4)
2024 ≤22,691 t as quoted in the agreed record (ICES’s 2025 sheet gives ≤15,378 t, the sum of its later substock advice) 31,301 t whole stock (agreed record, 13 Dec 2023)
2025 ≤15,511 t 25,028 t whole stock, “reflect[ing] … the ICES FMSY scenarios for the Northwestern and Viking sub-stocks” (agreed record, 6 Dec 2024)
2026 Zero catch for all substocks. An alternative without precautionary considerations for two substocks gave ≤12,280 t 14,034 t whole stock (11,164 t in Subarea 4), 44% below 2025 but above even the no-precaution alternative (agreed record, 22 Dec 2025)

Verdict: strengthened. The bias the chapter highlighted continued. Managers kept setting catches above closure advice, most recently a 14,034 t TAC for 2026 against zero-catch advice. The component-level collapse and the 2025 zero-catch advice bear out its warning. Its logical point that “has not yet crashed” is not evidence of resilience still stands.

Weight for the lessons. The lessons on hindsight-convergent estimates, “real-world conditions” (unaccounted removals) and institutional override can be weighted heavily. North Sea cod is now a second, independent case alongside northern cod.

Claim 5. EU/ICES precautionary reference points (B_lim at “immediate danger of collapse”; B_pa near the old MBAL, the minimum biologically acceptable level) were a “precautionary gloss” on unchanged practice, in conflict with the UN Fish Stocks Agreement standard that F_MSY is a minimum limit reference point; Member States appear unwilling or unable to fund recovery (pp. 23–24)#

Subsequent developments. - The UN standard is unchanged. Annex II para. 7 of the UN Fish Stocks Agreement still reads: “The fishing mortality rate which generates maximum sustainable yield should be regarded as a minimum standard for limit reference points.” Art. 6(2) says the absence of adequate scientific information “shall not be used as a reason for postponing or failing to take conservation and management measures” (https://www.un.org/depts/los/convention_agreements/texts/fish_stocks_agreement/CONF164_37.htm). - The 2002 CFP reform. Regulation 2371/2002 (20 Dec 2002) adopted a precautionary-approach definition along UNFSA lines and required “as a priority, recovery plans for fisheries exploiting stocks which are outside safe biological limits” (https://eur-lex.europa.eu/eli/reg/2002/2371/oj). - The 2013 CFP reform. Regulation 1380/2013, Art. 2(2), requires that “the maximum sustainable yield exploitation rate shall be achieved by 2015 where possible and, on a progressive, incremental basis at the latest by 2020 for all stocks” (https://eur-lex.europa.eu/eli/reg/2013/1380/oj). - ICES moved from managing to the MBAL to MSY-based advice. - ICES technical guidelines (1 March 2021) now define B_lim as the level “below which a stock is considered to have reduced reproductive capacity”. They set F_MSY so that the long-term probability of SSB falling below B_lim is ≤5%, within an advice rule that cuts F below a trigger level called MSY B_trigger (https://doi.org/10.17895/ices.advice.7891). - North Sea cod advice has been “MSY approach”-based since 2016 (ICES Advice 2025). - F_MSY is still used as a target, not a minimum limit in the UNFSA sense. - The 2020 deadline was missed. - The European Court of Auditors found “measureable improvement” in the Atlantic but “many stocks were still overfished”. It found Mediterranean fishing “at twice sustainable levels”, and that the Member States it visited used only 6% of their European Maritime and Fisheries Fund (EMFF) money on measures directly related to conservation. - It cited Pew’s finding that 42% of Northeast Atlantic TACs exceeded scientific advice in 2019 (ECA Special Report 26/2020: https://www.eca.europa.eu/Lists/ECADocuments/SR20_26/SR_Marine_environment_EN.pdf). - The EEA’s indicator (11 Aug 2026) says the MSY objective “has not yet been reached”. Only 40% of 235 assessed stocks meet both good-environmental-status (GES) criteria: 45% in the NE Atlantic and Baltic, 26.5% in the Mediterranean and Black Sea (https://www.eea.europa.eu/en/analysis/indicators/status-of-marine-fish-and). - Froese et al. (2018) found 69% of 397 European stocks subject to overfishing. They argued that management “which aims at maximum sustainable exploitation, is unable to rebuild the depleted stocks” (https://doi.org/10.1016/j.marpol.2018.04.018). - Unwillingness to fund recovery is visible in the TAC record. See the table of North Sea cod TACs set against closure advice under Claim 4.

Verdict: partly held up. The specific target of the critique (reference points built around the MBAL) was largely superseded by MSY-based advice and legally binding MSY objectives. That is a real change the chapter did not foresee in detail. The deeper critique still holds: - F_MSY is treated as a target, not a limit. - Decisions have repeatedly exceeded the advice. - The deadlines were missed.

Weight for the lessons. The lesson that “precautionary language can relabel unchanged practice” is moderately strong. The EU case shows the language, and then the law, changing faster than outcomes. The northern cod reference-point revision (Claim 2) is a separate example of how the choice of reference point determines what “precautionary” means.

Claim 6. Model-based precaution is challenged by the cod case (“Yet still it crashed”); “second-generation” precaution (error-resilient no-take zones, ecosystem-based management, fishers’ knowledge) “has considerable potential if implemented”; Georges Bank closures “appear to be producing results” (pp. 22, 25–26)#

Subsequent developments. - Model-based management has worked for many stocks. This cuts against the chapter’s implied pessimism. - Hilborn et al. (2020) found that for scientifically assessed stocks (about half of global catch), “on average, abundance is increasing and is at proposed target levels”. Poorly managed regions have “3-fold greater harvest rates and half the abundance” (https://doi.org/10.1073/pnas.1909726116). - The US had rebuilt 51 stocks since 2000 by the end of 2024 (NOAA Status of Stocks 2024: https://www.fisheries.noaa.gov/national/sustainable-fisheries/status-stocks-2024). - Murawski (2010), the lead author of the Georges Bank paper the chapter cites, found that of 24 depleted stocks whose formal plans cut fishing mortality, “all but one exhibited signs of recovery” (https://doi.org/10.1093/icesjms/fsq125). - Barents Sea (Northeast Arctic) cod, the contrast case the EEA added in 2013, rebuilt after Norway’s drastic cuts in 1989–90. Under a joint Norwegian–Russian harvest control rule, both total and spawning stock “have grown since 2006 and peaked in 2013” (MOSJ, Norwegian Polar Institute environmental monitoring, updated 22 Mar 2024: https://mosj.no/en/indikator/fauna/marine-fauna/stock-of-northeast-arctic-cod/). - That success has not lasted. The joint Norwegian–Russian scientists’ 2027 advice (26 June 2026) puts 2026 spawning biomass “below MSY Btrigger and Bpa, but above Blim”, and “the lowest since 2000”. Median recruitment has been below the long-term average “during the last decade”, and advised 2027 catch is 312,667 t (IMR–VNIRO 2026: https://www.hi.no/en/hi/nettrapporter/imr-vniro-en-2026-1). - Unlike the North Sea and Newfoundland assessments, this one reports minimal year-to-year revision since a 2021 benchmark. So a well-run, model-based rule rebuilt a cod stock quickly but has not protected it from a long run of poor recruitment. - It has repeatedly failed for cod in changing ecosystems. - Pershing et al. (2015) showed that rapid warming cut Gulf of Maine cod recruitment and raised mortality, and that “failure to recognize the impact of warming on cod contributed to overfishing” (https://doi.org/10.1126/science.aac9819). - Georges Bank and Gulf of Maine cod are both “overfished and subject to overfishing” (2021 assessment). Georges Bank has a rebuilding target of 2026 (NOAA, updated 11 Aug 2026: https://www.fisheries.noaa.gov/species/atlantic-cod). - The Georges Bank closures had mixed results. - Haddock on Georges Bank and in the Gulf of Maine is “not overfished” (2024 assessment; NOAA, updated 25 June 2026: https://www.fisheries.noaa.gov/species/haddock). - Sea scallops recovered after 1998–2001 closures combined with rotational area management (NOAA, updated 29 June 2026: https://www.fisheries.noaa.gov/species/atlantic-sea-scallop). - Cod did not recover. - In 2018 NMFS approved removal of the Closed Area I and Nantucket Lightship groundfish/habitat closure designations. It disapproved removal of Closed Area II (83 FR 15240, effective 9 April 2018: https://www.govinfo.gov/content/pkg/FR-2018-04-09/html/2018-06760.htm). - Walters and Maguire (1996), writing on the northern cod collapse, are commonly read as recommending large-scale closures to regulate exploitation directly (https://doi.org/10.1007/BF00182340; abstract not retrieved, so this characterisation is unverified). - Ecosystem-based management is adopted in policy but thin in practice. - Of more than 1,200 stocks reviewed, ecosystem drivers were built into tactical management for only 24: “while the ecosystem approach is highlighted in policy, key aspects of it tend yet not to be implemented” (Skern-Mauritzen et al. 2016: https://doi.org/10.1111/faf.12111). - NOAA adopted an ecosystem-based fisheries management (EBFM) policy and road map. Bringing ecosystem models into its processes “has faced procedural challenges in many jurisdictions” (Townsend et al. 2019: https://doi.org/10.3389/fmars.2019.00641). - Northern cod is now assessed with a “capelin-informed” ecosystem model (SAR 2026/030). This is a real move toward the chapter’s vision, but it remains dependent on a model (see Claim 2). - Stock structure has been recognised elsewhere. - ICES now treats northern shelf cod as three substocks and manages “to protect the weakest substock” (ICES Advice 2025). - The New England Fishery Management Council set catch limits for “four new Atlantic cod stock units” (10 Dec 2024: https://www.nefmc.org/news/groundfish-council-takes-final-action-on-framework-69-with-catch-limits-for-four-new-atlantic-cod-stock-units). - Fishers’ knowledge now enters formal assessments. - For northern cod, the bounds on unreported catch were “determined during discussions involving stakeholders present at past assessment meetings” (SAR 2024/049). - Harvesters’ union representatives sit on DFO’s peer-review meetings (SAR 2025/043 participant list).

Verdict: partly held up. The challenge to model-based precaution was right for cod under shifting productivity. It was too pessimistic as a general judgement. Second-generation tools produced some successes (haddock, scallops and substock-aware advice) but did not rescue cod. Implementation has been partial and at times reversed.

Weight for the lessons. The claim that “precaution run through biased models fails” carries strong weight for stocks whose productivity is shifting and moderate weight in general. The claim that “model-independent set-asides add resilience” stays suggestive, not demonstrated for the target species the chapter cared most about. The sub-stock lesson (Heincke’s herring races; p. 26) is strongly vindicated.

Claim 7. California sardine: state scientists’ mid-1920s warnings were overridden by the federal Bureau’s “wise use” stance and a 1939 political replacement of state experts; the catch was raised to “assist the war-effort” and the stock collapsed in 1942, with recovery only in the mid-1980s (pp. 19–20; Table 2.1, p. 27)#

Subsequent developments. - The institutional story is corroborated by a participant’s account. - Radovich (1982), a former California state biologist, describes state scientists recommending seasonal limits of 200,000 tons (1931, 1934) and 250,000 tons (1938). - He describes federal biologists who “looked for causes, other than fishing pressures”, with debates “often based on the same data”. - He describes the industry’s “delaying tactic of advocating or sponsoring more research”, including the 1947 Marine Research Committee. He cites MacCall’s estimate that holding the catch to about 250,000 t would have kept the fishery viable (CalCOFI Rep. 23: https://calcofi.org/downloads/publications/calcofireports/v23/Vol_23_Radovich.pdf). - Radovich does not mention the 1939 replacement of state experts. That detail rests on McEvoy alone and was not verified. - The dating is wrong. - Radovich reports that the fishery “collapsed to a low point in 1947”. Sardines were last landed in British Columbia in 1947–48, in Oregon and Washington in 1948–49, and in San Francisco Bay in 1951–52. The California moratorium came in 1967. - His catch table, read from a scanned copy, shows California landings still around 400,000–590,000 tons a season through the early 1940s. - The chapter’s “1942” (repeated in Table 2.1) appears to conflate the 1942 “no reason to be concerned” report with the collapse. - The EEA’s 2013 report reprinted the error unchanged (“1942 Continuous inaction leads to the collapse of the sardine stock”, Annex 2, Table A2.2, p. 703). - The attribution is now weighted toward the environment. - Lindegren et al. (2013) model sardine–anchovy dynamics back to 1661 and conclude that “the sardine collapse of the 1950s was largely unavoidable given poor recruitment conditions”. Fishing “modifies the dynamics” (https://doi.org/10.1073/pnas.1305733110). - McClatchie et al. (2017) find sardine “collapsed 29–40% of the time” before commercial fishing (https://doi.org/10.1002/2016GL071751). - Chavez et al. (2003) document roughly 50-year Pacific regime shifts between sardine and anchovy (https://doi.org/10.1126/science.1075880). - A second collapse repeated the pattern. - Zwolinski and Demer (2012) warned that “the repetition of the fishery’s response to a declining sardine stock”, meaning rising exploitation of the oldest fish, was “alarming”. They found productivity “below modeled estimates used to derive the current fishery-exploitation rates” (https://doi.org/10.1073/pnas.1113806109). - The directed fishery “was closed in 2015, and has not reopened”. The stock is “overfished”, with a rebuilding target of 2035 (NOAA, updated 31 Aug 2026: https://www.fisheries.noaa.gov/species/pacific-sardine).

Verdict: partly held up. The account of the institutional conflict is well corroborated. The collapse date is wrong. Table 2.1’s causal framing (“Continuous inaction leads to the collapse”) is weakened by later work that treats the environment as the primary driver and fishing as a factor making it worse. The second collapse, however, supports the chapter’s broader point that management kept exploitation high while productivity fell.

Weight for the lessons. The lessons on “political override” and split jurisdiction (state against federal) and on industry-sponsored delay can carry moderate-to-strong weight. Any lesson that treats the 1940s–50s collapse as simply caused by inaction should be weighted low. A better formulation: a fishery that did not reduce catches as a natural downturn set in turned an environmental decline into a longer and deeper collapse.

Claim 8. Scotland 1898–1998: landings (~333,000 t) and fleet tonnage (~109,000 t) stayed roughly constant while boats fell from 11,536 to 2,661 and fishers from 36,161 to 7,771; late-1990s net return on capital for the demersal fleet was 0.1%; North Sea industrial catches peaked in the 1970s at 2.2 of 3.5 Mt, presented as “fishing down the food web” (p. 19)#

Subsequent developments. - The trends continued. - Scottish vessels landed 532,000 t (£734M) in 2024, with 1,998 active vessels and 3,735 fishers (Scottish Sea Fisheries Statistics 2024, first published 27 Oct 2025, updated 27 Feb 2026: https://www.gov.scot/publications/scottish-sea-fisheries-statistics-2024/). - Landings are now dominated by pelagic species. The fisher count has roughly halved again since 1998. - Profitability remains marginal for the North Sea/West of Scotland demersal fleets. - In Seafish’s 2024 fleet economics (July 2025), the North Sea/West of Scotland demersal over-24m segment had operating margins of 10–12% but a net profit margin of −5% in 2023. Demersal seiners were at −6% and pair trawl/seine at 0% (Table 8: https://www.seafish.org/media/nffax2xh/amended-09-04-economics-of-the-fleet-2024.pdf). - Accounting definitions differ from those behind the 2001 figure, so the comparison is indicative only. - The 2001 Green Paper (Vol. 1) itself concluded that “The policy has not delivered sustainable exploitation of fisheries resources and will need to be changed if it is to do so” (COM(2001)135: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52001DC0135). The 0.1% figure from Vol. 2 was not re-verified. - “Fishing down the food web” is contested as a general indicator. - Essington et al. (2006) found the decline in mean trophic level usually reflects adding low-trophic fisheries (“fishing through”). But “only in the North Atlantic were ecosystems regularly subjected to sequential collapse and replacement of fisheries” (https://doi.org/10.1073/pnas.0510964103). - Branch et al. (2010) found that “catch MTL does not reliably predict changes in marine ecosystems” (https://doi.org/10.1038/nature09528). - Pinsky et al. (2011) found that collapses are, if anything, more frequent among small, low-trophic species (https://doi.org/10.1073/pnas.1015313108). - The North Sea, where demersal catches were supplanted by industrial ones, is the region where the sequential-replacement reading is best supported.

Verdict: partly held up. The long-run pattern of more capability without more yield, and the thin profitability, are borne out. The “fishing down” label is contested as a global metric, though it fits the North Atlantic relatively well. Two of the section’s figures could not be independently re-checked.

Weight for the lessons. “Capability masks decline” and “capability gains absorbed by depletion” keep moderate weight, since the evidence is endpoint comparisons. The trophic-level framing deserves low-to-moderate weight and needs its caveats stated.

Claim 9. Reducing fishing pressure would “greatly increase the economic returns”, but depleted natural capital means industry cannot absorb the short-term cost of rebuilding; the open question is whether change is “fast enough to stave off further collapses” (p. 26)#

Subsequent developments. - The economic claim is strongly supported by later work. - Costello et al. (2016) estimate that reforms could add more than 16 Mt of catch and “$53 billion in profit” a year, with “the median fishery taking under 10 y to reach recovery targets” (https://doi.org/10.1073/pnas.1520420113). - Froese et al. (2018) find that exploitation at 50–80% of the maximum would rebuild European stocks with “substantially higher profits” (https://doi.org/10.1016/j.marpol.2018.04.018). - Worm et al. (2009) report falling exploitation rates in 5 of 10 well-studied ecosystems, but 63% of assessed stocks “still require rebuilding” (https://doi.org/10.1126/science.1173146). - The “rebuilding trap” is also supported. - Neubauer et al. (2013): “prolonged intense overexploitation, especially for collapsed stocks, not only delays rebuilding but also substantially increases the uncertainty in recovery times” (https://doi.org/10.1126/science.1230441). - Murawski (2010): successful programmes used “substantial, measurable reductions in fishing mortality at the onset, rather than relying on incremental small reductions”. - Shelton et al. (2006): continued fishing under low productivity was “further delaying recovery”. - Subsidies, a point the chapter raised (p. 26), finally got a global rule. The WTO Agreement on Fisheries Subsidies was adopted on 17 June 2022 and entered into force on 15 Sept 2025 (https://www.wto.org/english/tratop_e/rulesneg_e/fish_e/fish_e.htm). - “Fast enough”? The answer is mixed. - In northern European waters, fishing mortality fell to sustainable levels on average (EEA 2026). - Globally, the share of stocks within biologically sustainable levels fell to 62.3% in 2021 (FAO SOFIA 2024, June 2024: https://www.fao.org/newsroom/detail/fao-report-global-fisheries-and-aquaculture-production-reaches-a-new-record-high/en). - For cod, which is what the chapter was mainly about, further collapses were not staved off: eastern Baltic, western Baltic and Kattegat cod, and the southern North Sea substock, all have zero-catch advice (Claim 4). Georges Bank and Gulf of Maine cod remain overfished (Claim 6).

Verdict: partly held up. The economics were right and have been strengthened by later modelling. The open prediction resolved unfavourably for North Atlantic cod and more favourably for many other assessed stocks.

Weight for the lessons. The rebuilding trap is a strong lesson. The claim that restraint pays is well supported by models, and more modestly by realised outcomes. Warming and ecosystem degradation, which the chapter barely mentions, now co-determine whether restraint is enough.

Claim 10. The burden of proof has shifted: 1998 US guidance requires protective measures “even in the absence of scientific certainty” (“a reversal of the burden of proof”); EPAP (1998) says uncertainty is no excuse to delay ecosystem-based management (p. 23 n. 3; p. 25)#

Subsequent developments. - The shift was codified in law. - The EU (2002 and 2013) defines the precautionary approach as one where “the absence of adequate scientific information should not justify postponing or failing to take management measures” (Reg. 1380/2013, Art. 4). - Canada’s amended Fisheries Act lets the Minister consider “the application of a precautionary approach and an ecosystem approach” (s. 2.5: https://laws-lois.justice.gc.ca/eng/acts/f-14/page-1.html). It requires rebuilding plans below the limit reference point (s. 6.2). - US performance under catch limits: “approximately 91 percent of all stocks or complexes did not exceed their annual catch limits”. At the end of 2024, 23 stocks were subject to overfishing and 42 were overfished (NOAA Status of Stocks 2024). - Scientific advice now often applies precaution as a default, for example ICES’s zero-catch advice “based on precautionary considerations” for northern shelf cod (ICES Advice 2025). - Implementation has been uneven or reversed. - TACs have been set above advice (Claims 4 and 5). - Canada reopened northern cod when projections showed a majority probability of decline at all catch levels examined (Claim 2). - US Executive Order 14276 (17 Apr 2025) directs the Commerce Secretary to “immediately consider suspending, revising, or rescinding regulations that overly burden America’s commercial fishing” and to review marine national monuments for opening to commercial fishing (https://www.whitehouse.gov/presidential-actions/2025/04/restoring-american-seafood-competitiveness/). - Ecosystem-based management remains rare in tactical decisions (Skern-Mauritzen et al. 2016).

Verdict: partly held up. The reversal of the burden of proof is real in law and in scientific advice. In decisions it holds only partly, and it has proved politically reversible.

Weight for the lessons. “Asymmetric proof standards make inaction the default” stays a strong descriptive lesson. That formal legal reversal of the burden of proof secures precautionary outcomes should carry only moderate weight. The recurring failure point is the step from advice to decision, not the drafting of principles.

Cross-cutting observations for the analytical lens (technology-neutral)#

  1. When a limit is redefined, “status” can change without the resource changing. The northern cod reclassification (Claim 2) and ICES’s successive revisions of reference points show that the thresholds defining “safe” are themselves model outputs. They can drift in the direction of optimism, and DFO says so explicitly. This extends the chapter’s point that what precaution protects is a value choice (p. 25).
  2. The gap between advice and decision is the most durable failure mode. It shows up in five places: - North Sea cod from 2001 to 2026 (including a 2026 TAC set against zero-catch advice); - the EU’s TAC record; - Harris in 1990; - the absence of the 1991 cut that Harris said would be “imperative” if fishing mortality stayed at 0.4 or above; - the northern cod decisions of 2024–26.

Barents Sea cod after 1989 is the main counter-example: advice was followed quickly, and the stock rebuilt. The chapter’s institutional lesson on “regulatory independence” (p. 26) is better supported now than in 2001. 3. Retrospective bias is systemic, and diagnosing it does not remove it. A standard statistic exists (Mohn’s ρ), but the bias still appears in 2025–26 assessments on both sides of the Atlantic. 4. Ecosystem and climate drivers complicate a simple overfishing story. Capelin, cold-water events, warming and hypoxia are now built into assessments. This is fairer to the chapter’s “uncertainty versus ignorance” lesson than to any single-cause narrative, including the chapter’s own treatment of the sardine collapse. 5. Transparency improved. DFO’s recent reports publish their own retrospective failures and the ways their reference points depend on models, which is what the chapter recommended (p. 22).

Limitations of this check#

Sources#

EEA 2013 report (how it revisited this chapter) - EEA (2013). Late lessons from early warnings: science, precaution, innovation, Annexes 1–3 (pp. 685–719), including Annex 2 (Tables A2.1–A2.2, pp. 702–703) and Annex 3, J. Ruzzin, “Fisheries/aquaculture: emerging challenges” (p. 718). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-annex-1/@@download/file - McGlade, J., van den Hove, S. (2013). “Ecosystems and managing the dynamics of change”, Ch. 17 in EEA Report 1/2013, pp. 407–428 (section 17.2.2 “A tale of two cods”; Panel 17.1). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2/late-lessons-chapters/late-lessons-ii-chapter-17/@@download/file

Northern cod: 1990 primary source and secondary accounts - Harris, L. (Feb 1990). Independent Review of the State of the Northern Cod Stock: Final Report. Northern Cod Review Panel, DFO. https://waves-vagues.dfo-mpo.gc.ca/library-bibliotheque/114276.pdf - Brubaker, E. (18 Jan 2000). “Unnatural Disaster: How Politics Destroyed Canada’s Atlantic Groundfisheries”, Environment Probe (secondary). https://environment.probeinternational.org/2000/01/18/unnatural-disaster-how-politics-destroyed-canadas-atlantic-groundfisheries/

Fisheries and Oceans Canada (DFO) and Government of Canada - DFO (Aug 2024). Northern (2J3KL) Atlantic Cod Assessment Framework. CSAS Sci. Advis. Rep. 2024/046. https://publications.gc.ca/collections/collection_2025/mpo-dfo/fs70-6/Fs70-6-2024-046-eng.pdf - DFO (Sept 2024). NAFO Divisions 2J3KL Northern Cod Stock Assessment to 2024. SAR 2024/049. https://publications.gc.ca/collections/collection_2025/mpo-dfo/fs70-6/Fs70-6-2024-049-eng.pdf - DFO (Sept 2025). NAFO Divisions 2J3KL Northern Cod Stock Assessment to 2025. SAR 2025/043. https://publications.gc.ca/collections/collection_2025/mpo-dfo/fs70-6/Fs70-6-2025-043-eng.pdf - DFO (July 2026). NAFO Divisions 2J3KL Northern Cod Stock Assessment to 2026. SAR 2026/030. https://publications.gc.ca/collections/collection_2026/mpo-dfo/fs70-6/Fs70-6-2026-030-eng.pdf - DFO (26 June 2024). 2024 2J3KL Northern cod fishery management approach. https://www.dfo-mpo.gc.ca/fisheries-peches/decisions/fm-2024-gp/atl-30-eng.html - DFO (18 June 2025). 2025 2J3KL Northern cod fishery management approach. https://www.dfo-mpo.gc.ca/fisheries-peches/decisions/fm-2025-gp/atl-20-eng.html - DFO (12 June 2026). 2026-27 2J3KL Northern cod fishery management approach. https://www.dfo-mpo.gc.ca/fisheries-peches/decisions/fm-2026-gp/atl-32-eng.html - DFO. Sustainability Survey for Fisheries – 2024 results (accessed 25 Sept 2026). https://www.dfo-mpo.gc.ca/reports-rapports/regs/sff-cpd/survey-sondage/index-en.html - Fisheries Act (R.S.C. 1985, c. F-14), ss. 2.5, 6.1, 6.2 (current consolidation, accessed 25 Sept 2026). https://laws-lois.justice.gc.ca/eng/acts/f-14/page-1.html ; https://laws-lois.justice.gc.ca/eng/acts/f-14/page-2.html - SOR/2022-73 (registered 4 Apr 2022; Canada Gazette Part II, 13 Apr 2022). https://gazette.gc.ca/rp-pr/p2/2022/2022-04-13/html/sor-dors73-eng.html - Heritage NL (Memorial University; undated). Economic Impacts of the Cod Moratorium (secondary). https://www.heritage.nf.ca/articles/economy/moratorium-impacts.php - Global Seafood Alliance (8 July 2024). Canada lifts 30-year cod fishing ban off Newfoundland and Labrador to mixed reactions (secondary). https://www.globalseafood.org/advocate/canada-lifts-30-year-cod-fishing-ban-off-newfoundland-and-labrador-to-mixed-reactions/

ICES - ICES Advice 2025, cod.27.46a7d20 (23 Sept 2025). https://doi.org/10.17895/ices.advice.27202566 - ICES Advice 2025, cod.27.22-24, western Baltic (28 May 2025; v2 4 Feb 2026). https://doi.org/10.17895/ices.advice.27202560 - ICES Advice 2024, cod.27.24-32, eastern Baltic (31 May 2024; v2 6 June 2024). https://doi.org/10.17895/ices.advice.25019216 - ICES Advice 2026, cod.27.21, Kattegat (30 June 2026). https://doi.org/10.17895/ices.advice.30932012 - ICES (2020). Workshop on Catch Forecast from Biased Assessments (WKFORBIAS; outputs from 2019 meeting). https://doi.org/10.17895/ices.pub.5997 - ICES (1 Mar 2021). Technical Guidelines – ICES fisheries management reference points for category 1 and 2 stocks. https://doi.org/10.17895/ices.advice.7891 - ICES Stock Assessment Graphs database (accessed 25 Sept 2026). North Sea cod assessments 2001 (key 102), 2011 (key 112), 2019 (key 13211); northern shelf cod substocks 2025 (keys 19661–19663). https://sag.ices.dk/SAG_API/api/SummaryTable?assessmentKey=102 (substitute key); reference points: https://sag.ices.dk/SAG_API/api/FishStockReferencePoints?assessmentKey=102

EU–Norway–UK catch-limit decisions and Northeast Arctic cod - Scottish Government (13 Dec 2023). EU, Norway and UK fisheries consultations: agreed records 2024, Cod. https://www.gov.scot/publications/european-union-norway-and-the-united-kingdom-fisheries-consultations-agreed-records-2024/pages/cod/ - Scottish Government (6 Dec 2024). EU, Norway and UK fisheries consultations: agreed records 2025, Cod. https://www.gov.scot/publications/european-union-norway-and-the-united-kingdom-fisheries-consultations-agreed-records-2025/pages/cod/ - Scottish Government (22 Dec 2025). EU, Norway and UK fisheries consultations: agreed records 2026, Cod (paras 15.1, 15.21). https://www.gov.scot/publications/european-union-norway-and-the-united-kingdom-fisheries-consultations-agreed-records-2026/pages/cod/ - European Commission (5 Dec 2025). The EU, Norway and the UK agree on fishing opportunities in the North Sea for 2026. https://oceans-and-fisheries.ec.europa.eu/news/eu-norway-and-uk-agree-fishing-opportunities-north-sea-2026-2025-12-05_en - Fishing News (16 Sept 2026). Single 2027 TAC for Northern Shelf cod likely (secondary). https://fishingnews.co.uk/news/single-2027-tac-northern-shelf-cod-likely/ - IMR–VNIRO (26 June 2026). Advice on fishing opportunities for Northeast Arctic cod in 2027. https://www.hi.no/en/hi/nettrapporter/imr-vniro-en-2026-1 - MOSJ (Environmental Monitoring of Svalbard and Jan Mayen; updated 22 Mar 2024). Stock of Northeast Arctic cod. https://mosj.no/en/indikator/fauna/marine-fauna/stock-of-northeast-arctic-cod/

European Union and EEA - Council Regulation (EC) No 2371/2002 (20 Dec 2002). https://eur-lex.europa.eu/eli/reg/2002/2371/oj - Regulation (EU) No 1380/2013 (11 Dec 2013). https://eur-lex.europa.eu/eli/reg/2013/1380/oj - European Commission (20 Mar 2001). Green Paper on the future of the CFP, COM(2001)135. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52001DC0135 - European Court of Auditors (2020). Special Report 26/2020, Marine environment: EU protection is wide but not deep. https://www.eca.europa.eu/Lists/ECADocuments/SR20_26/SR_Marine_environment_EN.pdf - EEA (11 Aug 2026). Indicator: Status of marine fish and shellfish stocks in European seas. https://www.eea.europa.eu/en/analysis/indicators/status-of-marine-fish-and - EEA (22 Jan 2013). Late lessons from early warnings: science, precaution, innovation, Summary (EEA Report 1/2013). https://www.eea.europa.eu/en/analysis/publications/late-lessons-2

United States - NOAA Fisheries. Atlantic cod (updated 11 Aug 2026). https://www.fisheries.noaa.gov/species/atlantic-cod - NOAA Fisheries. Haddock (updated 25 June 2026). https://www.fisheries.noaa.gov/species/haddock - NOAA Fisheries. Atlantic sea scallop (updated 29 June 2026). https://www.fisheries.noaa.gov/species/atlantic-sea-scallop - NOAA Fisheries. Pacific sardine (updated 31 Aug 2026). https://www.fisheries.noaa.gov/species/pacific-sardine - NOAA Fisheries. Status of Stocks 2024 (page updated 12 Aug 2026). https://www.fisheries.noaa.gov/national/sustainable-fisheries/status-stocks-2024 - NMFS (9 Apr 2018). Omnibus Essential Fish Habitat Amendment 2 final rule, 83 FR 15240. https://www.govinfo.gov/content/pkg/FR-2018-04-09/html/2018-06760.htm - NEFMC (10 Dec 2024). Council takes final action on Framework 69 with catch limits for four new Atlantic cod stock units. https://www.nefmc.org/news/groundfish-council-takes-final-action-on-framework-69-with-catch-limits-for-four-new-atlantic-cod-stock-units - Executive Order 14276 (17 Apr 2025). Restoring American Seafood Competitiveness. https://www.whitehouse.gov/presidential-actions/2025/04/restoring-american-seafood-competitiveness/

UK/Scotland - Scottish Government (27 Oct 2025; updated 27 Feb 2026). Scottish Sea Fisheries Statistics 2024. https://www.gov.scot/publications/scottish-sea-fisheries-statistics-2024/ - Seafish (July 2025). 2024 Economics of the UK Fishing Fleet. https://www.seafish.org/media/nffax2xh/amended-09-04-economics-of-the-fleet-2024.pdf

International - UN (1995). Agreement for the Implementation of the Provisions of UNCLOS … relating to Straddling Fish Stocks and Highly Migratory Fish Stocks, Art. 6(2), Annex II para. 7. https://www.un.org/depts/los/convention_agreements/texts/fish_stocks_agreement/CONF164_37.htm - FAO (June 2024). SOFIA 2024 news release. https://www.fao.org/newsroom/detail/fao-report-global-fisheries-and-aquaculture-production-reaches-a-new-record-high/en - WTO. Agreement on Fisheries Subsidies (adopted 17 June 2022; in force 15 Sept 2025). https://www.wto.org/english/tratop_e/rulesneg_e/fish_e/fish_e.htm

Peer-reviewed literature - Branch, T.A. et al. (2010). The trophic fingerprint of marine fisheries. Nature 468:431–435. https://doi.org/10.1038/nature09528 - Chavez, F.P. et al. (2003). From anchovies to sardines and back. Science 299:217–221. https://doi.org/10.1126/science.1075880 - Cook, R., Sinclair, A., Stefánsson, G. (1997). Potential collapse of North Sea cod stocks. Nature 385:521–522. https://doi.org/10.1038/385521a0 - Costello, C. et al. (2016). Global fishery prospects under contrasting management regimes. PNAS 113:5125–5129. https://doi.org/10.1073/pnas.1520420113 - Essington, T.E., Beaudreau, A.H., Wiedenmann, J. (2006). Fishing through marine food webs. PNAS 103:3171–3175. https://doi.org/10.1073/pnas.0510964103 - Frank, K.T. et al. (2005). Trophic cascades in a formerly cod-dominated ecosystem. Science 308:1621–1623. https://doi.org/10.1126/science.1113075 - Froese, R. et al. (2018). Status and rebuilding of European fisheries. Marine Policy 93:159–170. https://doi.org/10.1016/j.marpol.2018.04.018 - Hilborn, R. et al. (2020). Effective fisheries management instrumental in improving fish stock status. PNAS 117:2218–2224. https://doi.org/10.1073/pnas.1909726116 - Horwood, J., O’Brien, C., Darby, C. (2006). North Sea cod recovery? ICES J. Mar. Sci. 63:961–968. https://doi.org/10.1016/j.icesjms.2006.05.001 - Hurtado-Ferro, F. et al. (2015). Looking in the rear-view mirror. ICES J. Mar. Sci. 72:99–110. https://doi.org/10.1093/icesjms/fsu198 - Hutchings, J.A., Myers, R.A. (1994). What can be learned from the collapse of a renewable resource? Can. J. Fish. Aquat. Sci. 51:2126–2146. https://doi.org/10.1139/f94-214 - Lindegren, M. et al. (2013). Climate, fishing, and fluctuations of sardine and anchovy in the California Current. PNAS 110:13672–13677. https://doi.org/10.1073/pnas.1305733110 - McClatchie, S. et al. (2017). Collapse and recovery of forage fish populations prior to commercial exploitation. Geophys. Res. Lett. 44:1877–1885. https://doi.org/10.1002/2016GL071751 - Mohn, R. (1999). The retrospective problem in sequential population analysis. ICES J. Mar. Sci. 56:473–488. https://doi.org/10.1006/jmsc.1999.0481 - Murawski, S.A. (2010). Rebuilding depleted fish stocks: the good, the bad, and, mostly, the ugly. ICES J. Mar. Sci. 67:1830–1840. https://doi.org/10.1093/icesjms/fsq125 - Myers, R.A., Hutchings, J.A., Barrowman, N.J. (1997). Why do fish stocks collapse? Ecol. Appl. 7:91–106. https://doi.org/10.1890/1051-0761(1997)007[0091:WDFSCT]2.0.CO;2 - Neubauer, P. et al. (2013). Resilience and recovery of overexploited marine populations. Science 340:347–349. https://doi.org/10.1126/science.1230441 - Pauly, D. et al. (1998). Fishing down marine food webs. Science 279:860–863. https://doi.org/10.1126/science.279.5352.860 - Pershing, A.J. et al. (2015). Slow adaptation in the face of rapid warming leads to collapse of the Gulf of Maine cod fishery. Science 350:809–812. https://doi.org/10.1126/science.aac9819 - Pinsky, M.L. et al. (2011). Unexpected patterns of fisheries collapse in the world’s oceans. PNAS 108:8317–8322. https://doi.org/10.1073/pnas.1015313108 - Radovich, J. (1982). The collapse of the California sardine fishery: what have we learned? CalCOFI Rep. 23:56–78. https://calcofi.org/downloads/publications/calcofireports/v23/Vol_23_Radovich.pdf - Rose, G.A. (2004). Reconciling overfishing and climate change with stock dynamics of Atlantic cod over 500 years. Can. J. Fish. Aquat. Sci. 61:1553–1557. https://doi.org/10.1139/f04-173 - Rose, G.A., Rowe, S. (2015). Northern cod comeback. Can. J. Fish. Aquat. Sci. 72:1789–1798. https://doi.org/10.1139/cjfas-2015-0346 - Rose, G.A., Walters, C.J. (2019). The state of Canada’s iconic Northern cod: a second opinion. Fisheries Research 219:105314. https://doi.org/10.1016/j.fishres.2019.105314 - Schijns, R. et al. (2021). Five centuries of cod catches in Eastern Canada. ICES J. Mar. Sci. 78:2675–2683. https://doi.org/10.1093/icesjms/fsab153 - Shelton, P.A. et al. (2006). Fishing under low productivity conditions is further delaying recovery of Northwest Atlantic cod. Can. J. Fish. Aquat. Sci. 63:235–238. https://doi.org/10.1139/f05-253 - Skern-Mauritzen, M. et al. (2016). Ecosystem processes are rarely included in tactical fisheries management. Fish and Fisheries 17:165–175. https://doi.org/10.1111/faf.12111 - Townsend, H.M. et al. (2019). Progress on implementing ecosystem-based fisheries management in the United States. Front. Mar. Sci. 6:641. https://doi.org/10.3389/fmars.2019.00641 - Walters, C.J., Maguire, J.-J. (1996). Lessons for stock assessment from the northern cod collapse. Rev. Fish Biol. Fish. 6:125–137. https://doi.org/10.1007/BF00182340 (abstract not retrieved) - Worm, B. et al. (2009). Rebuilding global fisheries. Science 325:578–585. https://doi.org/10.1126/science.1173146 - Zwolinski, J.P., Demer, D.A. (2012). A cold oceanographic regime with high exploitation rates in the Northeast Pacific forecasts a collapse of the sardine stock. PNAS 109:4175–4180. https://doi.org/10.1073/pnas.1113806109