Aligning stock assessment and harvest strategies in data-limited, multi-species fisheries: A hierarchical modeling approach to BC's flatfish fishery

Contemporary fisheries management involves setting quantitative objectives, and establishing decision rules that determine management actions in response to monitoring data. Within that system, feedback harvest strategies are developed via simulation, where various decision rule features are tested against realistic fishery conditions, and are ranked by their relative performance measured against the quantitative fishery objectives. While the adoption of formal harvest strategies has been increasing globally for some time, most contemporary harvest strategies, and their associated stock assessments, continue to be single-species oriented despite a high prevalence of technical interactions in fisheries. Over three research chapters, I use a simulation approach to investigate whether multi-species harvest strategies, based on hierarchical stock assessment models and multi-species reference points that incorporate technical interactions among species, are better suited to the management of data-limited, multi-species fisheries than traditional single-species approaches. First, I use simulation-evaluation to show that the estimation performance of hierarchical multi-stock surplus production models is more robust to declining data quality than a single-stock version of the same model, creating potential for improved information feedbacks in data-limited contexts. Next, I show that TACs set based on hierarchical model estimates of biomass and productivity are better able to maximise yield under all data quantity scenarios, thanks to negatively correlated biases in key management parameters and target harvest rates that acknowledge technical interactions. Finally, I use closed loop simulation to compare the economic and conservation risks of single-species maximum sustainable yield (MSY), multi-species MSY, or maximum economic yield (MEY) harvest strategies. Taken together, the results of my thesis support expanded usage of hierarchical models in fisheries stock assessment and management, which may lead to wider adoption of formal harvest strategies. Further, I show that yields, food security, and/or economic rent can be increased (not always at the same time) by including technical interactions in reference point calculations.