LAUSR

Fishery catch projection models play a central role in fishery management, yet are underrepresented in the literature. A wide range of statistical approaches are employed for the task, including multiple regression models, autoregressive methods, different classes of generalized linear models, mixed model approaches and many others. However, the applicability of these statistical approaches can be limited in specific cases of complex fisheries. We developed a new catch projection model for quota-based fisheries on the West Coast of the U.S. to forecast annual catch and landings for a variety of groundfish species in the Northeast Pacific Ocean. The model projects total and landed catch of each species by individual vessel and for the entire fishing fleet, using a combination of weighted mean vessel attainment rates and historical catch rates, and generates uncertainty intervals. It demonstrated an ability to produce highly accurate predictions at both fleet (R2=0.9847) and vessel levels (R2=0.8447). The model framework contains much built-in versatility, is generalizable enough to serve a variety of quota based applications, and the approach can be tailored to other fisheries around the world. With the proliferation of quota based management of commercial fisheries, tools such as this one are increasingly useful for sustainable management of fishery resources.