Incidental mortality of dolphins in the eastern Pacific Ocean purse-seine fishery: correlates and their spatial association

Abstract

A zero-inflated Poisson model was used to identify typical fishing practices that contributed to incidental mortality of dolphins in the eastern Pacific purse-seine fishery between 1993 and 2001. The presence of hazardous net conditions (net canopies and net collapses), the duration of the backdown procedure (the primary method of releasing dolphins from the net), the size and species composition of the encircled dolphin herd and the amount of tuna encircled, were all found to consistently contribute to increased dolphin mortality per set. In particular, the presence of net canopies and large biomass in the net contributed to both the development of problematic situations in which mortality could occur and to the mean mortality per set, once a problematic situation had developed. On the other hand, lengthy backdown procedures and the presence of net collapses contributed to the development of problematic situations, but had less effect on the mean mortality per set once a problematic situation had developed. Because some of these variables are partially correlated, the overall conclusion of this analysis is that one of the primary causes of dolphin mortality continues to be the encirclement of large herds. Dolphin mortality can increase with the number of dolphins encircled because: (1) the more animals encircled, the greater the likelihood of entanglement and mortality while confined in the net; and (2) the duration of the backdown procedure increases with the number of animals encircled. The duration of the backdown procedure may, in turn, contribute to increased dolphin mortality by: (1) keeping dolphins in close contact with the net for longer periods of time, thereby increasing the chances for entanglement; and (2) leading to the formation of net canopies. Dolphin mortality increases in the presence of net canopies because animals can be trapped below the sea surface in the areas of canopies. Spatial distributions of encircled herd size, duration of the backdown procedure, presence of net canopies and presence of dolphin mortality show similar patterns. Encircled herd size tended to be greatest south of the equator and north of the equator along the offshore margin of the fishery. In these areas, the duration of the backdown procedure tended to be longer and there was often an increased probability of net canopies and dolphin mortality, but also larger catches of tuna. These consistent spatial patterns suggest that reallocation of fishing effort to other areas may be an effective means of reducing the current level of dolphin mortality. Predictive models could be developed to assess tradeoffs between dolphin mortality and tuna catches at varying levels of fishing effort in areas where large herds are targeted by fishermen and different strategies for reallocation of fishing effort to other areas or to purse-seine sets on unassociated tunas.