Since 1960, landings of Atlantic herring have been the greatest of any marine species in Canada, surpassing Atlantic cod and accounting for 24% of the total seafood harvested in Atlantic Canada. The Scotian Shelf-Bay of Fundy herring fisheries (NAFO Division 4VWX) is among Canada’s oldest and drives this productivity, accounting for up to 75% of the total herring catch in some years. The stocks' productivity and overall health have declined since 1965. Despite management measures to promote recovery implemented since 2003, biomass remains low and is declining. The factors that drive the productivity of 4VWX herring are primarily unresolved, likely impeding the effectiveness of management actions on this stock. We evaluated potential drivers of herring variability by analyzing 52 time-series that describe the temporal and spatial evolution of the 4VWX herring population and the physical, ecological, and anthropogenic factors that could affect them using structural equation models. Variation in herring biomass was best accounted for by the exploitation rate's negative effect and the geographic distribution of fishing and recruitment. Thermal phenology and temperature adversely and egg predation positively impacted the early life stage mortality rate and, ultimately, adult biomass. These findings are broadly relevant to fisheries management, but particularly for 4VWX herring, where the current management approach does not consider their early life stage dynamics or assess them within the ecosystem or climate change contexts.

We assembled data that included spatial and temporal variation in NAFO Division 4VWX herring across different life stages, the exploitation dynamics, and the biological and environmental factors that may influence herring. Data sources related to changes in the mean states, seasonal dynamics (timing, amplitude), and community composition of the plankton or larval assemblages that herring interact with, the type and intensity of fishing pressure, predation and competition were obtained or calculated from the stock assessments, peer-reviewed publications, at-sea surveys, ships of opportunity, and remote sensing are listed in Table 1. From these data, 52 time series were developed (Table 2).  Five of the series describe the temporal variability in larval, juvenile, and adult herring dynamics; the remaining 47 describe the environmental, biological, and anthropogenic factors that could drive them. A full description of the data sources and methods to calculate these indices are in the Supplementary Information (SI). The indices are organized within eight categories (Table 2): 1) Intrinsic (Herring; n=5); 2) Anthropogenic (n=5); 3) Competition (n=1); 4) Predation (n=6); 5) Plankton (n=4); 6) Physical (n=22); 7) Prey (n=2); 8) Phenology (n=7). Both raw and imputed datasets are provided.