How bycatch rates of ETP species vary spatiotemporally and with covariates

Since 2010, in Denmark, electronic monitoring (EM) has been routinely done in gillnet fisheries to collect data on bycatch of endangered, threatened and protected (ETP) species. The EM data is collected at a fine spatial and temporal scale, including time, position, and effort of each fishing operation, whether or not there was a bycatch event. Effort is in the form of net-length and soak-time. The EM data allowed us to model bycatch per unit effort and how it changed with location, time, and different aspects of the fishery. Separately, standard national logbook data gives us an overall approximation of the total effort in the Danish gillnet fisheries. These data were then used to extrapolate estimates of annual total bycatch of ETP species. Annual total bycatch is a nonlinear function of the model parameters including a spatio-temporal random effect. Recent statistical developments have shown that, in this case, estimates can be improved using a bias correction method that was recently implemented in glmmTMB. Here, we present these statistical models and the resulting estimates of annual bycatch of cormorants, common eiders, and alcides in Danish gillnet fisheries.