Fact-checking fisheries citations: where the primary texts are reachable (FAO SOFIA 2026, TP 721, Costello 2008)
Direct PDFs work with a browser User-Agent: fao.org/3/cd8357en/cd8357en.pdf (SOFIA 2026) and fao.org/3/cd5538en/cd5538en.pdf (Technical Paper 721); pdftotext -layout then grep. openknowledge.fao.org DSpace pid API returned non-JSON. iisd.org and sciencedirect.com return 403 to curl; an IISD statement was readable through a republication on impakter.com. A Science reprint of Costello et al. 2008 is on forest-trends.org. Gotchas: SOFIA 2026 gives 62.4 percent of assessed stocks sustainable in 2023, so 37.6 percent overfished is a subtraction, and FAO attributes the drop from 64.5 partly to methodology. The 72.6 percent volume figure applies to landings of assessed stocks, not all global landings. Costello note 20 excludes territorial and community concessions, so it is weak evidence for community-based rights. Frank et al. 2011 Nature reports the Scotian Shelf moving back toward benthic fish dominance, which dates claims resting on Frank 2005.
Two refinements after reading the sources. (1) Costello note 20 opens 'Other forms of property rights may induce similar incentives' and excludes territorial/community concessions because of spatial scale relative to LME catch data, so it is a limit of the sample rather than evidence against them. Costello also ran propensity-score matching, though the 'about half' figure itself is the raw share collapsed by 2003. (2) For cod-specific status on the eastern Scotian Shelf, DFO Science Advisory Report 2011/028 is more direct than Frank 2011: 4VsW spawning biomass about 7,500 t in 2003, 64,000 t in 2009, with projections of renewed decline under 1994-2009 productivity. I found no later 4VsW assessment. Europe PMC REST returns abstracts for paywalled Science/Nature papers.