Investigating coastal ecosystem structure and dynamics using Ecopath mass-balance modelling and stable isotope data

Ecosystem modelling and stable isotope analysis are both widely applied in marine ecology. Both methods have been used to study food web structure and predator prey interactions. Stable isotopes have also been used to compare the trophic level from the Ecopath model with the trophic level from the nitrogen isotope as a validation of the Ecopath model, and the main aim of this PhD thesis was to investigate whether we can use stable isotope analysis to depict diet data to use as input to an Ecopath mass balance ecosystem model. The study area has been Ullsfjord-Sørfjord, a northern fjord ecosystem at 70° N. Ullsfjord is the outer part of the system and is larger, more open towards the western Barents Sea and therefore highly influenced by Atlantic water. Sørfjord is the inner part of the system and is smaller, shallower, colder in winter and has lower species diversity of birds, fish and semipelagic crustaceans. The impact of fishing is very low in Sørfjord. We wanted to use as much local data as possible in the Ecopath model, and conducted extensive field work, sampling and collecting data on biomass and diet for as many of the groups in our model as possible. We show that stable isotopes can be used to detect niche shifts. In this case for haddock in Ullsfjord we showed how haddock feeding habits at a young stage (10 to 25 cm in length) differed from those of larger haddock and from small and large cod. Haddock had a pelagic feeding isotope signature and more pelagic than what has been reported in literature. We also showed that several of the most common fish species in Ullsfjord had overlapping carbon isotope signature at sizes between 25 and 45 cm, indicating they may be feeding on similar prey and suggest trophic redundancy. We investigated structure and flow in the two ecosystems Ullsfjord and Sørfjord, using data from 1993 – 96 and Ecopath models of the two fjords. We found that while Sørfjord was dominated by cod as a top predator and keystone species, Ullsfjord, with its higher species diversity, showed a more web-like structure and several lower trophic level groups were keystone groups from the Ecopath model. For the final part and main objective of this project we used a MixSIR mixing model to estimate possible prey contributions from the stable isotope signatures. This was used to change the diet matrix of an Ecopath model from 2010 based on gut content analysis (GCA)data. Then the two models (GCA and stable isotope) were compared. Graphical abstract of main objective of this thesis is shown in fig. 1. We conclude that stable isotopes and mixing models show good promise for being both an alternative way of getting diet information,especially on lower trophic levels, and an addition to GCA on higher trophic levels to reduce uncertainty and increase robustness in the model.