How Observation Coverage Shapes Bycatch Metrics in the Tropical Tuna Purse Seine Fishery

Investigating the effects of fishing on non-targeted (bycatch) species is crucial for conservation and management matters. On-board observers information provides a valuable source for estimating bycatch; however, due to several factors such as their high costs and logistic challenges, observer programmes usually cover a small percentage of the fishery. In order to estimate bycatch for the unobserved fishing activity, a ratio estimator is commonly used, which assumes a linear relationship between the ratio of bycatch and total target catch or effort. In this study, we implemented a simulation experiment to evaluate the performance of the ratio and modelbased estimator under different sampling coverage scenarios. We used the Spanish tuna purse seine fishery operating in the Atlantic Ocean as a case study. Our results suggest that the ratio estimator may produce bycatch estimates with large negative bias (i.e., underestimation) when the sampling coverage is lower than 20%, even for common taxa. Conversely, the model-based estimator produced unbiased estimates even under low sampling coverage scenarios. However, the model-based estimator may be only suitable for taxa with intermediate and high prevalence in the bycatch composition. This study presents a simulation framework that may be applied to other moderately data-rich fisheries and supports the implementation of observer programmes from which appropriate estimates of bycatch for species of interest to the Commission can be obtained.