A Spatially-Explicit Operating Model to Assess the Impact of Varying Tag Experimental Designs on the Performance of a Brownie Tagging Model for Atlantic Yellowfin Tuna

A spatially-explicit, individual-based, capture-recapture model was constructed to simulate alternative experimental designs for the proposed Atlantic Yellowfin Tuna tagging program.  Releases were modeled either as continual (to mimic an opportunistic tagging program) or discrete events.  After accounting for handling mortality and tag shedding, the population of tagged fish was simulated using a spatial population dynamics model that allowed migration amongst four geographic regions including North America, South America, West Africa, and a combined Europe-North Africa region.  Tag recoveries were from the major commercial fisheries and calculated based upon fishing mortality and observer coverage (i.e., reporting rate).  A Brownie model was then implemented to predict regional tag recoveries by fitting to the simulated observations and estimating parameters (i.e., natural mortality, fishing mortality, movement, and fishery selectivity by fleet).  Multiple facets of the tagging program were investigated to determine the best protocols for tag releases (e.g., release frequency, numbers, and age sampling valuation) and recapture design (e.g., observer coverage, fisheries reporting, or genetic sampling) under a variety of assumed population dynamics (e.g., mortality and movement rates).  The simulation-estimation framework allows appraisal of the cost-benefit of alternative experimental designs, while evaluating tagging data utility for informing stock assessment parameters of highly migratory species.