Integrating key ecosystem components into stock assessment models for fishery management

Small pelagic fishes, particularly European anchovy (Engraulis encrasicolus) and sardine (Sardina pilchardus), are fundamental to the ecology and fisheries of the Adriatic Sea. Their short life cycles, mid-trophic position and strong sensitivity to environmental variability make them highly responsive to the combined effects of fishing pressure and climate change. Over recent decades, both species have experienced persistent declines in mean body size and productivity, challenging stock resilience and exposing limitations in assessment approaches that assume stationary population dynamics. This thesis examines how fishing and environmental forcing interact to shape long-term trends in size structure and population dynamics, with the objective of improving the ecological realism of stock assessments. A multidisciplinary framework was adopted, integrating Generalized Additive Models (GAMs) and Stock Synthesis (SS). GAMs were used to disentangle the relative influence of fishing activity, cumulative temporal effects and key environmental drivers on size dynamics, while SS models explicitly embedded environmental covariates into population dynamics through time-varying biological parameters and survey catchability. Results consistently indicate that sustained fishing pressure and its long-term legacy are the dominant drivers of size decline, with temporal signals capturing the cumulative effects of historical overexploitation and limited recovery capacity. Environmental variability plays a secondary role, modulating species-specific growth, recruitment and catchability rather than acting as an independent decline driver Incorporating environmental forcing within SS improves coherence between observed biological patterns and modeled dynamics, reducing the risk of biased estimates of stock status and reference points under changing environmental conditions. Overall, this work demonstrates that combining flexible statistical approaches with mechanistic stock assessment models provides a robust pathway toward ecosystem-informed fisheries management. By aligning assessment outputs more closely with ecological processes, this framework supports adaptive management of environmentally sensitive, short-lived species and offers transferable insights for other marine systems facing the joint pressures of exploitation and climate change.