Estimating dead fish quantities dropping out of gillnets when direct observations are impossible

Abstract Fish that are caught in passive or fixed fishing gear and subsequently die and drop out of the gear, or that are removed by scavengers, do not show up in catch statistics and constitute unaccounted-for removals. Estimating these removals is challenging, particularly when direct observations are not possible because fishing conditions preclude the use of cameras or other devices or means to catch or observe them. Using a simple theoretical process model, we define two independent modelling approaches to estimate lost catch resulting from drop-out. The first is based on a minimally refined analysis of commonly available fishery catch per unit effort data. The second model uses data that can be obtained from field experiments and from at-sea fishery observers, specifically data on catch amounts, and the composition of the catch relative to fish condition—live, dead undegraded, and dead degraded. We apply these modelling approaches to data from the Gulf of St Lawrence (Canada) Greenland halibut (Reinhardtius hippoglossoides) gillnet fishery, for which multi-day soak times have long been suspected of generating significant lost catch while only little evidence is available. In fact, available evidence naively contradicts this suspicion, as the observed proportion of catches that are degraded and ultimately discarded in this fishery is typically small overall and increases only a little with increased soak duration. The two modelling approaches produce similar results, estimating that total removals of Greenland halibut were on average 4.7 to 5.4 times the recorded landings over the period from 2000 to 2024. This result is consistent with the high mortality levels suggested by the assessment for this stock. Of broader relevance to other gillnet fisheries worldwide that may employ shorter soak time, we found that estimated unobserved catch losses equalled retained catch after soak times of ~15 h. Importantly the condition of fish in catches may belie the quantities of dead fish losses. Failing to account for fishing derived loss of the magnitude estimated here could result in important biases in stock assessments and associated sustainable harvesting frameworks, in addition to constituting an important source of waste.