Exploring food-web dynamics. The Non-Deterministic Network Dynamics model (NDND) is a simulation model designed to explore the possible dynamics of the Barents Sea food-web. Typically, the NDND produces multiple outputs that represent different possible trajectories of the food-web.

Major challenges. Many plant and animal groups are not well observed or are not part of monitoring programs. There are large uncertainties about the relationships between prey and predators. The diet of many animals is also known to vary between individuals, between regions and between seasons and years, but we have little information to describe precisely these variations. All of these make it difficult to build realistic simulation models for real marine food-webs.

A model based on Chance and Necessity. The NDND model recognizes the limitations and uncertainties in our knowledge and is constructed on a limited set of assumptions and constraints that reflect this limited knowledge. In the NDND the biomass fluctuations of the different species are determined by the balance between gains (=feeding) and losses (=predation, metabolic losses, fishing). Unlike other food-web models, the NDND does not specify feeding in a deterministic manner. Rather, consumption of available prey is left to “chance and necessity”, where chance reflects indeterminacy and necessity reflects ecological constraints, such as prey availability or maximum feeding and growth rates.

Important questions. The outputs of the NDND models are sets of possible food-web dynamics, under specific assumptions. We can use the NDND model to answer questions like:

• Could there be food-web configurations in the Barents Sea that are different from those we have observed so far?
• Are there common/average food web configurations and rare/extreme ones? What would trigger changes from one type to the other?
• How far into the future can we expect to be able to predict the state of the Barents Sea food-web?
• Can changes in the ocean climate (temperature) or fishing regime lead to different types of food-web dynamics?

Be prepared for changes. We have used the NDND model to explore what could be the possible state of the Barents Sea food web, beyond what wehave observed in the past few decades. What we found is that there is a large potential for change and reorganization of the food-web. It is possiblethat earlier food-web structures were quite different from those we observe today. It is equally possible that future food-web structures will be different from anything we have observed in the past. Small pelagic fish species, like capelin and herring, are particularly prone to large «swings» in biomass, with consequences that propagates across the entire ecosystem. More at https://doi.org/10.1371/journal.pone.0254015