Incorporating Uncertainty into Fishery Models

Coping with Uncertainty: Evolution of the Relationship between Science and Management

Pamela M. Mace and Michael P. Sissenwine

doi: https://doi.org/10.47886/9781888569315.ch2

Abstract.— Scientists, managers and fishers have always known that uncertainty is an integral part of nature. Early assessment models dealt with uncertainty as a signal-noise problem, developing deterministic, equilibrium relationships to explain cause and effect. Other models coped with uncertainty by normalizing model outputs against inputs that were known to be highly variable (for example, yield per recruit analysis, which avoids consideration of recruitment variability by performing calculations on a per recruit basis). Although deterministic approaches are still widely used, it is also true that methods for quantifying and presenting uncertainty have been taught, developed, and applied for several decades. The question has been less one of how to present uncertainty, rather than whether to present it. What purpose would it serve? Experience showed that managers or politicians presented with a range of estimates frequently chose to set quotas from the risk-prone end of the range. In short, uncertainty was used as a reason for avoiding conservation measures. Development of the precautionary approach to fisheries management has initiated a transition in the perception of uncertainty from that of a reason to avoid action to that of a reason to exercise caution. Increased public awareness of the limits to the productivity of natural marine resources and increasing numbers of economically devastating stock collapses have been instrumental in initiating the transition, more so than scientific advances in the modeling and representation of uncertainty. This transition is, however, more active on paper than in actual implementation. Today, one challenge is to integrate science and management into a rigorous, risk-averse, decision analysis framework, and another is to further promote responsible management and the application of the precautionary approach in the face of uncertainty. Presentation of uncertainty as risk profiles or Bayesian posterior distributions rather than confidence intervals around point estimates has provided a more objective basis for evaluating the consequences of alternative management decisions, but there is a need for better comprehension of the full consequences of the large risks managers, fishers and other stakeholders (which includes the public in general) often appear to be willing to take.