Kill, Kill, Kill! : Simulating Marine Ecosystems without Dogfish (or Cod, or Herring, or Birds)

Large marine ecosystems are complicated, with ecological components that are highly connected both to each other (via trophic interactions) and their environment.  As ecosystem-based fishery management (EBFM) moves from questions of why one would do EBFM towards how to do EBFM, recognition of the importance of these connections has increased.  Studies have shown that commercially important fish stocks may migrate out of regional areas and management units, but little research has explored the likely effects on the biotic community following such a shift.  Additionally it has been hypothesized that reasons for the lack of recovery of some commercially important species, despite significant management actions, could include predation by or competition with other, less commercially desirable species.  Again, little has been done to explore what the effects of removals of these less desirable species might have on the species of interest, a community, or even an entire ecosystem.  To that end we submit that fishery-oriented ecosystem models are a powerful way to explore such questions.  Scenarios where individual species or groups are removed from a system, thus simulating either migration out of an area or targeted removals to achieve a management goal, can elucidate some of the probable community or ecosystem responses following such a removal.  Doing such experiments in silico experiments, particularly when using multi-model inference, can also allow fishery managers to explore the potential effects on a management unit when a species is removed from the community due to climate change or fishing pressure.  We present results from models parameterized for the Northeast United States large marine ecosystem which explore such removals.  These models include a full system model, ATLANTIS and a multi-species surplus production model which incorporates species interactions, MS-PROD.  Our results indicate that removals of species or groups from an ecosystem are unlikely to have simple effects due to the complexity of the interactions between species and their environment.  The propensity for indirect effects and unanticipated consequences is high in these scenarios; as such we recommend the use of such ecosystem models to explore the full range of options germane for implementing EBFM.