MEPS 447:231-242 (2012)  -  DOI: https://doi.org/10.3354/meps09449

ABSTRACT: Fisheries observer data were used to model the distribution of seabird encounters in order to identify potential areas for fisheries closures. Data from the Australian Eastern Tuna and Billfish Fishery (ETBF) in the Tasman Sea were used, and the focal species was the flesh-footed shearwater Puffinus carneipes Gould, 1844 (the species most commonly killed in this fishery). Encounters between flesh-footed shearwaters and longline fishing vessels varied with season, ­distance from Lord Howe Island, and a number of environmental and oceanographic variables. Encounters were most common south-west of Lord Howe Island, in waters associated with the Tasman Front and the East Australian Current. The resulting model was used to predict overlaps between fisheries and flesh-footed shearwaters during 3 years (1997-98, 2003-04 and 2006-07). During 2003-04, high rates of interaction were predicted in areas with high fishing effort and high observed mortality rates of shearwaters. In 2006-07 most fishing was well to the north of areas with predictions of high interaction rates, and seabird bycatch was low. The shift in fishing ­locations coincided, and was likely driven by, a change in the fish species targeted by the majority of the fishery. This indicates that the most likely reason for falling bycatch rates in this fishery were movements of the fishing effort away from the birds, rather than changes in fishing ­technique. These results emphasise the potential of area closures as a method of bycatch mitigation for ­species that are proving intractable to standard bycatch reduction methods. Our results demonstrate that data collected from fisheries vessels can be used to identify characteristic areas of interactions.

KEY WORDS: Bootstrap model averaging · Long-line fishing · By-catch · Lord Howe Island · ­At-sea interactions