| Summary: | Multispecies interactions (predation and competition) are known to have important consequences for the dynamics of marine fish populations. These interactions depend on the spatial overlap among fish species in the community. Several approaches have been used to quantify species interactions, including production models and age- (or length) structured multispecies models. In this study, multi-species biomass dynamics models were extended to account for food-web interactions in multiple spatial areas (Gulf of Maine, Southern New England, and Georges Bank). A total of 15 fish species collected from the study areas were aggregated into four trophic groups: non-migrating benthivores (haddock, yellowtail flounder, winter flounder, and little skate), non-migrating piscivores (Atlantic cod and summer flounder), migrating piscivores (silver hake, spiny dogfish, winter skate, goosefish, pollock, and white hake), and migrating planktivores (Atlantic herring, Atlantic mackerel, longfin squid). The spatial distribution of each species group was determined from trawl-survey data, taking into account distributional shifts. We assumed that the migratory groups (planktivores and piscivores) range over the entire study area, such that their production can be described with a single set of model parameters (r and k). By contrast, production of non-migrating groups (piscivores and benthivores) was assessed with a different set of model parameters (r and k) for each spatial area. A hierarchical model fitting procedure was used to estimate the production parameters (r and k) and interaction coefficients among migrating and non-migrating species groups. In our study, migrating groups (F and P) played a spatially essential role in species interactions across multiple areas, indicating that the three spatial areas are functionally connected through the high degree of connectivity and direct linkages between migrating groups (F and P) and non-migrating groups (B and S). Our results demonstrate that accounting for trophic interactions improves the model fit and that the strength and direction of these interactions vary among spatial areas. Based on the area-specific interaction effects, this approach can help us understand the functional connections among multiple areas and thus inform current fisheries management.
|
|---|