A Poisson and negative binomial regression model of sea turtle interactions in Hawaii's longline fishery

Abstract

Sea turtle interactions with the longline fishery in Hawaii have become a serious concern in recent years. Various measures, including a swordfish harvest ban, have been adopted to protect sea turtles. This study explores the factors and degree of sea turtle interactions with Hawaii's pelagic longline fishery, i.e. the production of undesirable outputs in the pursuit of an economic activity, in an analytical framework of rare events using the count data models for the period 1994-2003. The analysis was based on the type of trip, such as those targeting tuna or swordfish. The fishing technologies associated with the choice of trip or target species, season, and turtle population explained turtle interactions with the longline fishery. The impact of a fishing vessel's previous history of turtle interactions had a negligible impact on subsequent interactions. In the absence of new longline fishing technologies to dramatically avert sea turtle interactions, there are about 6% and 55% chance that at least one turtle per trip may be encountered in tuna-and swordfish-targeted fishing trips, respectively. This study confirms that more turtle interactions are associated with the swordfish-targeted trips. Several of the factors that contributed substantially to turtle interactions can be feasibly regulated. Furthermore, the conservation and management of sea turtles require increased policy dialogues and cooperation among the coastal nations vis-à-vis the adaptation to "turtle-friendly" fishing technologies.

Driftnet fishery threats sea turtles in the Atlantic Ocean

“…This pattern holds in several different fisheries, such as pelagic longlines (Gilman et al 2006;Pradhan and Leung 2006;Sales et al 2010), drift gillnets (Gallaway 2001) and trammel nets (Cambiè 2011). The capture of sea turtles is rare (but see a different situation in Alfaro-Shigueto et al 2011 for small-scale fisheries).…”

Quantitative evaluation of the performance of a permit auction system in reducing bycatch of sea turtles in the Hawaii Swordfish longline fishery

“…We make use of estimates of turtle mortality (k ¼ 0.61) and fishery profits n ¼ 0.17 based on Pradhan and Leung [1].…”

“…Hence, we have modeled sea turtle encounters in the Hawaii swordfish fishery as events governed by a Poisson process with intensity l (i.e., in a very short time period dt, the probability of an encounter with a sea turtle is l dt and for periods that do not overlap with each other, the probability of sea turtle encounter is independent). To estimate the value of l, we minimized the differences between the probabilities of sea turtle encounters generated by the Poisson process and the observed probabilities in Pradhan and Leung [1], essentially employing a non-linear least squares optimization procedure.…”

“…According to Pradhan and Leung [1], during the period 1991-2001 there were on average 125 longliners working the swordfish fishery off Hawaii. Here we assume that 125 longliners in total participate in the fishery, and that they are owned by five fishing firms.…”

By-catch in the Saginaw Bay, Lake Huron commercial trap net fishery

“…Literature suggests that the negative binomial is often a reasonable probability distribution for analyzing catch data because these data often exhibit over-dispersion (Claramunt et al, 2009;Clark, 1974;Power and Moser, 1999;Pradhan and Leung, 2006;Ward and Myers, 2005). Additionally, preliminary analysis that also involved testing the Poisson and geometric distributions indicated that assuming the negative binomial probability distribution provided the best fit to the observed data.…”