Uncertainty in the management of activities affecting marine mammal populations. The tuna-dolphin conflict, a case study

Model performance and observation error were acknowledged and formally included into an assessment about the status of the NE stock of spotted dolphins that is affected by the tuna fishery in the ETP. The results of the assessment were then used to evaluate alternative future scenarios that could occur under different management goals and the application of contrasting policies. In general, models with a higher level of complexity did not provide better estimates of parameters. The level of population depletion was estimated reasonably well by all models. For decision parameters if the allowable bycatch was low, there was no resulting bias induced risk to make erroneous predictions about population status. If the bycatch was large, there were occasions where the population status and trend were incorrectly determined. Results of a sensitivity analysis agreed with a revision of the indices of relative abundance that rendered inappropriate for population assessments. This situation may be leading to the false perception that the dolphin stock is not recovering despite the substantial reductions in the bycatch. The population model included a component to estimate additional mortality attributed to the fishery, unrecorded by the observers. The Bayesian posterior distributions for the parameter that controls this unobserved mortality (UOM) have most of their density near zero. The resulting estimates of UOM with highest probabilities were in the low hundreds of animals. The conclusion from this part of the analysis is that the UOM is low and not preventing the population from recovery, and therefore the population is actually increasing. A risk assessment indicated that under the current level of bycatch, the population would be increasing with different probabilities depending on the time a policy is allowed to act. The estimation of parameters of population dynamics, unobserved mortality and the risk assessment will need to be repeated as soon as the currently ongoing revision of the estimates of absolute abundance is finished.