Re-examining top-down forcing on the Eastern Scotian Shelf fish community using a multispecies size spectrum model

Fish community abundance and size structure consistently change under heavy fishing pressure. Efforts to rebuild fish populations and communities by reducing fishing are not always successful, in part due to indirect food-web effects, including the replacement of fishing mortality by that from top predators. On the Eastern Scotian Shelf in the northwest Atlantic, there has been limited recovery of depleted fish populations following the 1993 moratoria on directed Atlantic cod and haddock fisheries. One hypothesis for this limited recovery is increasing predation by growing grey seal populations, although the contribution of this top-down forcing remains uncertain. To explore the extent to which increasing grey seal predation may have contributed to observed fish community changes, we developed a multispecies size-spectrum (mizer) model and compared a factual scenario representing the historical changes in fishing and increase in seal predation with a counterfactual scenario without an increase in seal predation, across a large ensemble of randomized food webs. While both scenarios broadly captured observed catch trends across most fish species, species’ biomass trends diverged between scenarios, with the factual scenario generally showing closer agreement with observations than the counterfactual scenario. Although both scenarios reproduced the persistent nonrecovery of Atlantic cod, the factual scenario better matched its observed biomass trend. Increasing seal-associated predation exerted additional downward pressure on cod, but across the range of food-web configurations, this marginal effect weakened as forage-fish predation on small cod increased, consistent with cultivation–depensation dynamics. Species-level differences in modelled mean individual body weight over the hindcast period fell within the range of observations, but species-specific interannual changes in size structure were poorly captured. Both scenarios’ modelled size spectra broadly reflected the observed patterns, but the factual scenario produced more pronounced, nonlinear shifts across the fish community relative to the reference period, with small-size classes disproportionately affected. Refining the model with additional data and integrating bottom-up processes and other potential sources of forcing would enhance its practical value for application in the management strategy evaluation process.