Estimating species interactions in the Georges Bank fish community

Accounting for the interactions among fished species is a key component of the ecosystem approach to fisheries management.  The goal of this research is to incorporate species interactions into a statistical multispecies catch-at-age model.  This model incorporates commercial catch, fishery-independent survey catch, and food habits data to estimate predation parameters, rates of fishing, initial abundances and annual recruitment.  Multiple single-species models were first fitted to provide initial parameter estimates for single-species parameters and try to resolve inconsistencies between commercial and survey catch data.  These single-species models were then linked through predation, permitting the estimation of predation mortality and therefore the separation of predation from other sources of natural mortality.  The multispecies model was fitted to simulated data to determine which parameters were estimable and indicated confounding between predator size-preference parameters.  A subset of the Georges Bank food habits data was used to estimate these size-preference coefficients because prey lengths and ages could not be routinely measured.  We are currently finalizing the model for a subset of the Georges Bank fish community: Atlantic cod, silver hake and Atlantic herring.  By estimating species interactions, we aim to explicitly account for the effect of predation mortality on the time dynamics of natural mortality.