A comparison of methods to spatially represent pelagic longline fishing effort in catch and bycatch studies

Bycatch in fisheries has been recognized as a threat to many endangered populations of sea turtles, sea birds and marine mammals. Interactions between pelagic longline fisheries and critically endangered populations of leatherback sea turtles (Dermochelys coriacea) have led to temporary closures of the Hawaiian pelagic longline swordfish fishery and severe bycatch quotas. The negative impact of these events on both the populations of certain endangered species and the economic livelihood of the fishermen has resulted in a strong push from all sides to better understand bycatch events. Typically, analyses of longline catch and bycatch have examined fishing effort summarized over large areas (≥1°). Although aggregation of effort to this level may be necessary to account for uncertainty, confidentiality concerns, or to make comparisons across regions, it specifically limits the researcher's ability to characterize the local oceanographic factors that may drive individual bycatch events. Higher resolution analyses must be undertaken to identify such features. However, for these higher resolution analyses, the methods currently used to spatially represent pelagic longline fishing effort may significantly affect researcher's results. Here, we look at different methods to represent this fishing effort (i.e., points, centroids, polylines and polygons) at various resolutions (2 km to 5°) to better understand which method and spatial resolution are most appropriate. Our results validate the use of point features to represent fishing effort in previous low resolution studies of the Hawaiian pelagic longline fishery by showing that the set point method is suitable for studies with resolutions lower than 15 km. However, at higher resolutions (≤15 km) and in areas with more sparsely distributed fishing, aggregated effort values differed significantly between spatial representation methods. We demonstrate that the use of polygons to describe pelagic longline fishing effort is more representative and necessary for such high resolution analyses.