Latent Gaussian models to predict historical bycatch

13 Knowledge about how many sh that have been killed due to bycatch is an important 14 aspect of ensuring a sustainable ecosystem and shery. We introduce a Bayesian spatio15 temporal prediction method for historical bycatch that incorporates two sources of available 16 data sets, shery data and survey data. The model used assumes that occurrence of bycatch 17 can be described as a log-linear combination of covariates and random e ects modeled as 18 Gaussian elds. Integrated Nested Laplace Approximations (INLA) is used for fast calcula19 tions. The method introduced is general, and is applied on bycatch of juvenile cod (Gadus 20 morhua) in the Barents Sea shrimp (Pandalus borealis) shery. In this shery we compare 21 our prediction method with the well known ratio and e ort methods, and make a strong 22 case that the Bayesian spatio-temporal method produces more reliable historical bycatch 23 predictions compared to existing methods. 24