LAUSR

Abstract The main Hawaiian islands Deep 7 bottomfish fishery is a multispecies fishery that uses hook-and-line gear to catch seven primary species and is currently managed and assessed as a single population. Management plans support, and recent assessment reviews recommend, assessing species on an individual basis. When fisheries catch multiple species, but information about abundance is desired for a single species, it is important to use data subsets that reliably inform abundance trends for that individual species. We use a previously unanalyzed dataset collected from a subset of Deep 7 bottomfish fishermen who reported targeting practices –including the species or group of species targeted, fishing practices, catches, and environmental variables– while fishing in the summer of 2016. Using three statistical classification methods, we show that targeting occurs, and is primarily for two groups: opakapaka (Pristipomoides filamentosus), and a combined group of ehu (Etelis carbunculus) and onaga (Etelis coruscans). Depth was a significant explanatory variable for targeting, statistically distinguishing shallower opakapaka targeting from deeper ehu/onaga targeting at 188 m. In the absence of information about depth, catch composition data can be used to distinguish between opakapaka and ehu/onaga groups nearly as well as when depth is included. Applying these methods to the full fisher reporting database revealed that an opakapaka group and an ehu/onaga group could be distinguished. Further information on targeting within the fisher reported data, or consistent information on depth if targeting preferences cannot be reported, would improve classification. Ultimately, any of the methods used herein can be applied to filter fishing events based on predicted targeting to produce more representative indices of abundance for use in single-species stock assessments.