WP2.1 Conceptualise a Baltic food web model: We will construct a conceptual food web model including characteristic benthic and pelagic food chains for each of the three main basins, i.e. Bothnian Bay, Baltic proper and Kattegat. The construction of such a model will use a top-down approach, i.e. starting with avian and mammalian top predators followed by identifying their prey species of key ecological and commercial relevance, along the different consumer and producer niches. The selected species will maximise interlinking different food chains as well as the spatiotemporal potential. A tentative conceptual food web model is shown in Figure 1, but will be subject to the output of WP1.1. The final conceptual food web model will be used to prioritise species and food chains to be studied in WP2.2-3, as well as WP3-6.
Figure 1. Conceptual food web model for the Bothnian Bay.
WP2.2 Fill relevant data gaps on trophic interactions and energy cycling: We will use the final conceptual food web model (WP2.1) to quantify predator-specific diet compositions by taxonomic determination of prey remains, pellets and stomach content, as well as their isotopic tracing and DNA-barcoding. The latter two techniques will be used on biobanked or collected samples as well. Several isotopic tracers, including bulk tissue, amino acid and fatty acid-specific stable carbon, nitrogen and sulphur isotopes, will be quantified using recently optimised gas chromatography–combustion interfaced–isotope ratio mass spectrometry, an online system allowing for amino and fatty acid-profiling. DNA-barcoding will utilise next-generation sequencing of selected metabarcodes for fish (e.g. 16S), invertebrate (e.g. CO1), and phytoplankton species (e.g. 16S rDNA). Caloric contents will be determined by standard calorimetric analyses. The generated data will support the food web models (WP2.3-4) as well as WP3-6. We will organise an inter-laboratory study to guarantee the reliability of the analytical output of the different labs involved.
WP2.3 Model trophic interactions and their spatiotemporal dynamics: We will employ Bayesian mixing models specifically developed for bulk tissue stable isotopes, and optimise them for the above novel isotopic tracers. Multivariate, i.e. factor and cluster analysis, methods will allow for amino- and fatty acid-profiling and DNA-barcoding. Information-theoretic selection of Bayesian and univariate, linear non-linear and piece-wise models will allow investigating spatiotemporal dynamics. Reliability of the models will be tested using convergence analysis and statistical bootstrapping. The output of this WP will feed into WP2.4 and WP3-6.
WP2.4 Model food web energy cycling and their spatiotemporal dynamics: We will construct a mathematical ecological network analysis of energy cycling, using a similar top-down approach as mentioned above (WP2.1), using estimates of total energy requirements of species-specific stocks. The output of this WP will feed into WP6.
