Compositional forecasting of Chinook Salmon Evolutionarily Significant Units in bycatch for Pacific Hake fisheries

Abstract

Bycatch impacts on non-target species present significant management problems in diverse fisheries throughout the world. Despite successful efforts to minimize bycatch in US West Coast Pacific Hake fisheries, these impacts remain a concern, particularly for sensitive populations of Chinook Salmon. NOAA Fisheries needed predictive models to estimate proportions of Chinook Salmon Evolutionarily Significant Units (ESUs) expected in bycatch. We used genetic mixture analysis to estimate ESU proportions from at-sea bycatch between 2008 and 2015. Using latitude as a predictor and applying jackknife cross validation, we found Dirichlet regression more accurately estimated abundant ESUs, whereas multinomial logistic regression performed better with rare ESUs. This targeted, ESU-specific approach showed the spatial distribution of sensitive stocks in bycatch and supported NOAAs obligations to forecast impacts on listed ESUs. The overarching goal of this continuing work is to maximize sustainable harvest while protecting threatened and endangered Chinook Salmon ESUs.