Predation by marine mammals explains recent trends in natural mortality of Pacific Herring (<i>Clupea pallasii</i>) and changes expectations for future biomass

Abstract The recovery of marine mammals from historical over-exploitation in the 1970s represents one of the largest changes in trophic structure in the northeast Pacific Ocean over the last century, for which the impacts on key forage species such as Pacific Herring (Clupea pallasii) are poorly understood. This has prompted hypotheses that increasing marine mammal populations are the primary cause for productivity declines for some fish stocks and their lack of recovery to historical abundance levels. In this study, we evaluate such a hypothesis for Pacific Herring by quantifying historical predation rates by key predators, including cetaceans (Pacific Humpbacks, Grey Whales), pinnipeds (Steller Sea Lions, Harbour Seals), and piscivorous fish (Pacific Hake). Predation mortality is quantified via a novel approach that integrates a single-species catch-at-age model with estimates of predator consumption derived from bioenergetic models. We found that predator consumption, largely driven by Humpback Whales, explained increasing Pacific Herring natural mortality rates in recent years and could be used to forecast future mortality. Incorporating higher future natural mortality rates produced higher estimates of current stock status (1.09–1.2$B_0$) based on lower estimates of equilibrium unfished biomass (17.5–20.3 kt). Conversely, models that assumed mortality was more like the historical average had lower stock status (0.63$B_0$) and higher estimates of unfished biomass (32.4 kt). We demonstrate a practical approach for ecosystem modelling that can be used to develop operating model scenarios for management strategy evaluation, improving scientific defensibility by removing an element of analyst choice for future mortality scenarios. We discuss how simpler modifications to single-species model assumptions can be more pragmatic for providing fisheries management advice, while more complex multi-species or ecosystem models might provide more nuanced insights for exploring research questions related to multi-species ecosystems and fisheries interactions.