Résumé
Here we present an annual abundance index for silky shark for the period 2012 to 2021 based on standardized silky shark catch per FOB fishing set for the French tropical tuna purse seine fishery in the Indian Ocean. The methodology used general additive models with mixed effects (GAMMs) in a novel approach using three submodels, including one model for tropical tuna catch per set, and two others that form the components of a Δ log-normal model (i.e., a presence-absence model and an abundance when present model) for silky shark catch per set with total tuna catch used as predictors in those models. Results indicate an overall increasing trend in silky shark abundance over the study time period, though when interannual variability in catch of target tunas is removed model CPUE predictions are found to be more or less stable over time. Though there is evidence to support a potential increasing trend in abundance, there are a number of reasons to believe that these predictions may be overly optimistic, including low data coverage and a biased spatial distribution of fishing effort and observer coverage of bycatch in the original two years of the data (2012-2013) and poorly understood impacts of the implementation of a quota on yellowfin tuna catch in 2017.