In this study we aim at improving our understanding about the species interactions and the dynamics of the Norwegian Sea ecosystem by developing and using modern methods for identifying and quantifying the diet of these ecologically and economically important pelagic fish populations.
After gathering this crucial diet information with high temporal and spatial resolution, we will use modern statistical methods to scrutinize these data, and end-to-end ecosystem models (Atlantis and NORWECOM.E2E) to test the hypotheses regarding ecosystem dynamics from plankton, via fish, to sea birds under the current climate change.
The project work is divided into four work packages:
WP1: Development of methods for increasing the resolution and understanding of dietary interactions
This work package has two aims:
1) To develop tools for high throughput genetic identification of pelagic prey in the stomach contents of pelagic fish and ichthyoplankton.
2) To establish methods for developing a library of fatty acid and stable isotope libraries for the pelagic species.
WP2: Mapping the multispecies diet interactions in the Norwegian Sea
The aim of this work package is to increase the spatial and temporal resolution of dietary information and build time series using modern methods developed in WP 1. The main effort is collecting data form the research surveys in the Nordic Seas, and analyzing it using visual, genetic, and fatty acid/stable isotope methods.
WP3: Modelling the Ecosystem dynamics of the Norwegian Sea
This work package addresses estimation of the annual consumption of the various prey species by pelagic fish and seabirds with the help of two end-to-end ecosystem models (Atlantis and NORWECOM.E2E). Another aim is to quantify how varying fish abundance will affect intra- and interspecific competition for pelagic fish expressed through changes in individual growth. Simulations will also be run with present and future climatic conditions, to estimate how feeding conditions and species interactions can change due to global warming. Ensemble runs with the two models will give us a measure of the uncertainties in the models.
This work package will pull together all the results from WPs 1-3. An important part of this WP is the work of the PhD-candidate on methods for spatial and temporal correlation in ecological data. Our observations are mainly of binary type, for instance whether a particular species is present or absent at a spatial site. The probability of being present varies spatially, and can be modelled using a Gaussian random field. State-of-the-art methods and software for Gaussian latent random fields will be used in the project. In a multi species setting correlation across species can be modelled using multiple random fields.