Uncertainty in the Management of Activities Affecting Marine Mammal Populations; The Tuna-Dolphin Conflict, a Case Study.

Natural resource management has traditionally assumed that model parameters are known without error or that they are well described by point estimates. However, our capacity to obtain accurate data from the field is limited and models are usually more realistic with at least observation error associated to the parameters.

In the case of marine mammals, problems such as incidental mortality (bycatch) and directed take generally require decisions regarding the maximum allowable number of individuals that can be removed from a population while achieving different kinds of management goals. In any case, it is always of interest to know the status of a stock in the future compared with some reference point given different management regimes. This ratio is considered itself a derived parameter and is the target of many assessments.

Few instances exist where uncertainties are incorporated into the models used to establish limits on allowable removals from marine mammal populations. The two best known examples are Potential Biological Removal (PBR), a method used by the US Government, and the revised management procedure (RMP) of the International Whaling Commission. Both are used to set mortality limits. Although PBR considers the uncertainty in observed mortalities and estimates of abundance, the logistic parameters were assumed to be known without error. Sensibility analysis were conducted using only few alternative fixed parameter values. The RMP of the IWC on the other hand is a complete procedure that accounts for uncertainties in model parameters. Both PBR and RMP have been developed however, as a management tool that does not produce an output for predictions about the future in terms of a derived parameter such as the one previously described.

The goal of this studyl is to develop a tool that helps to define management decisions in terms of derived parameters that describe the potential situation of the stock of interest in the future. Such a method will incorporate the observation error associated to parameter estimates of a logistic model and different versions of an age structured model. The particular objectives of this project are: 1) To test the performance of different models for the particular conditions of the data sets of the dolphin stocks in the ETP and select a model to use for parameter estimation and prediction of abundance and trend; 2) To conduct estimation of population parameters for ETP dolphins using the most complete data set available; 3) To test the hypothesis that it is not necessary to use information beyond that provided by the dolphin biology and the reported mortality to explain the actual estimated trend in population abundance. The alternative hypothesis is that the recovery of dolphin stocks is being slowed down by some unobserved mortality attributable to the tuna fishery; 4) To conduct a risk assessment based on the calculation of the probabilities associated with future population status of ETP dolphins depending on different allowable bycatch levels.

A draft for the first chapter of my dissertation has been prepared and is titled: Model performance in the estimation of parameters of population dynamics and management of the north eastern stock of spotted dolphins (Stenella attenuata) affected by the tuna fishery in the eastern tropical pacific. Work on the second and third chapters is underway. Data was obtained from the Inter-American Tropical Tuna Commission (IATTC) and is being prepared for analysis. This consists of the estimation of parameters of population dynamics, including an estimate of an indirect effect of the fishery on the affected dolphin populations that is not being recorded by the observer program of the IATTC. The final chapter of my dissertation also depends on the recently obtained data and focuses on the evaluation of future management regimes and evaluation of the implicit risks.