In November 2015 we posted commentary from Desmond Kahn reacting to the 2014 IUCN listing of American eel, Anguilla rostrate, as an endangered species. He suggested to first evaluate the IUCN’s listing as a hypothesis, one which warranted an evaluation, rather than accept it as a final assessment. The brief review he offered in 2015 concluded that this endangered listing may not have been supported by existing data after all.
Taking a look at American eel data
Kahn expanded on this subject in his recently published paper (2019) where he developed an index of relative abundance by estimating the annual mean total catch of eels per recreational fishing trip and combined that with commercial landings of eels during the same time period (1981-2014). This methodology produced a different set of results and conclusions than the 2012 American eel stock assessment from the Atlantic States Marine Fisheries Commission (ASMFC), which was the primary citation in the IUCN’s decision to list this species as endangered.
Beginning in 1981, Kahn found an erratic drop in the index of relative abundance (recreational eel catch per trip) to a low of about one sixth of the original level by the late 1990’s. But from 2003 – 2014 catch per trip increased, such that in 2014 they were over half of the 1981-1982 levels. Looking at the commercial side, landings have been quite stable over the last two decades.
American eel abundance might instead be more closely linked to environmental factors like variances in the North Atlantic Oscillation (NAO) which has been associated with European eel abundance in other studies.
Issues with the 2012 stock assessment
The catch per recreational trip index developed to estimate the trends in abundance in Kahn (2019) were obtained from estimates of the total recreational catch and the total quantity of recreational fishing trips along the coasts of the United States. These parameters are estimated by the National Marine Fisheries Service in an integrated survey encompassing the entire Atlantic coast. Due to very large sample sizes, precision of the catch-per-trip index is relatively high. The 2012 assessment encountered difficulties because it attempted to integrate a large number of disparate state-specific and waterbody-specific fishery-independent surveys. The assessment team employed generalized linear modeling to estimate trends of twenty, thirty and forty years’ length, but the resulting trends contradicted each other, and the assessment failed to provide statistical justification for its model choices, a fairly common failing among stock assessments.
Recreational catch data was included in the 2012 assessment only in very low- precision estimates of the landings in weight; these misrepresented the increasing trend in total catch, including discards, since 2003, because of the high and increasing proportion of eels that were discarded, up to 90% by 2014. Since the assessment examined only recreational landings, and not total catch, the assessment team may have received a false impression of declining abundance in more recent years.
The stock assessment included virtually no data from the habitat where most maturing females occur, according to the literature, which is non-tidal freshwater, except for possibly some active eel weirs in freshwater nontidal rivers in New York and Maine. When females mature, they begin migration downstream to the Sargasso Sea, and reduce feeding, so they become less likely to enter commercial pots which are fished in estuaries, where fishery samples are largely obtained; consequently, female spawning stock biomass is probably heavily underrepresented in the data employed in the stock assessment, leading to a possible bias.
Additionally, the peer reviewers of the 2012 assessment concluded that, based on their criticisms of the depletion-based stock reduction model, the “…overfishing and overfished status …cannot be stated with confidence,” and the assessment was enjoined from drawing conclusions on those parameters.
Kahn (2019) also discusses the difficulty of estimating the natural mortality rates of American Eels, which were employed in parts of the assessment, due to the semelparous life history of the species. To estimate the trend in commercial fishing mortality, he avoided the need to use estimates of natural mortality by employing relative fishing mortality and found this mortality was no higher during the period of decline (1981-1995) than during the period of recovery (2003-2014), which hurt the hypothesis that increased fishing mortality had contributed to the decline.
A fundamental assumption for declaring a species endangered is that the species should be experiencing a recent decline or would presently be at historically low abundance. Kahn’s results challenge that notion and might have exposed a serious oversight by the IUCN.