Multiple Models for Marine Fisheries Management: What Fisheries Science Can Learn from Other Disciplines

In providing in providing model-based information for fisheries management, uncertainty inevitably rears its ugly head.  This uncertainty is in part attributable to uncertainty in our understanding of the underlying ecological processes and how to best quantify those processes (structural uncertainty). Additionally, uncertainty in the input parameter estimates results in uncertainty in the outputs of the model (parameter uncertainty).To deal with these sources of uncertainty, fisheries modelers are beginning to use multiple models and model ensemble approaches. To better understand how to use these approaches, NOAA-National Marine Fisheries Service held a workshop to hear from scientist in other disciplines (e.g., oceanography, climate science, weather) on methods and implications of multiple-model approaches. The invited presenters noted that, in their disciplines, multiple-model approaches are used regularly and have resulted in considerable improvement in model projections and forecasts (e.g., reduction in the uncertainty in estimates of hurricane strength and trajectory, reduction in the uncertainty of climate projections, and reduction in the uncertainty of ocean circulation patterns).  These model upgrades have improved the advice given to a wide range of stakeholders including emergency planners, the IPCC community of stakeholders, and oil spill response teams.