Incorporating predator-prey information in fisheries stock assessments is challenging due to the multitude, complexity, and dynamic nature of trophic interactions. However, integrating indices of abundance derived from predator diet data into assessment models as proxy abundance indices of poorly sampled age classes may have utility as an ecosystem approach that improves assessment outcomes. Here, we leveraged 34 years of observations of common terns (Sterna hirundo) provisioning age-1 Atlantic herring (Clupea harengus) to their chicks across 12 islands in the Gulf of Maine to construct an annual index of herring provisioning rates. We then integrated this index into a state-space, age-structured assessment model as an index for age-1 Atlantic herring (recruit) abundance. Our results showed that when the assessment model was fit assuming a linear relationship between the herring provisioning index and recruitment, the model produced a residual pattern indicating a non-linear response between the provisioning of Atlantic herring and their abundance in the environment. These findings suggested that Atlantic herring provisioning rates plateaued during periods of high recruitment (satiation) and remained consistent at low levels during periods of low recruitment (hyperstability). Allowing for this non-linearity in the stock assessment improved the fit to the index. Ultimately, the seabird diet index may be a useful indicator to incorporate into the Atlantic herring assessment process; however, the quantitative incorporation of this index into the assessment model is challenged by non-linear predator-prey dynamics. |
