Precautionary design of a marine protected area based on a habitat model

Abstract

A basic principle of effective resource management is that decisions should be con- servative in the face of uncertainty. Due to limited data, there is often considerable uncertainty about species' habitat relationships and requirements. If the boundaries of a protected area are based on relationships estimated by a habitat model, effective management takes the uncertainty into account. The inclusion of uncertainty in the design of a hypothetical marine protected area is described for a coastal population of the long-beaked common dolphin Delphinus capensis off Baja California, Mexico. Line-transect and depth data were combined in a hierarchical Bayesian model. Two possible management goals were considered: protecting 100 000 animals or protect- ing 60% of the population. A precautionary approach was adopted, meaning that the manage- ment goal should be met with a high probability. The model estimated that a seaward boundary at 360 m would include 100 000 dolphins with a probability of 0.9. A conventional but less precau- tionary 'best estimate' boundary at 160 m would meet the management goal with a probability of 0.5. For the second goal of including 60% of the population, the precautionary and non-precau- tionary depths were 210 and 170 m, respectively. Habitat models are useful for management, but management decisions based on such models should consider the uncertainty inherent in estimat- ing parameters from data. Models which include the data observation process can improve infer- ence about habitat relationships.