Diet uncertainty analysis strengthens model-derived indicators of food web structure and function

Abstract Ecological Network Analysis (ENA) can inform marine management decisions by producing indicators that describe ecosystem health and function. Reporting ENA indicators with uncertainty boundaries lets end-users draw stronger inferences and can increase confidence in model results. However, few studies developing these indicators have estimated uncertainty due to data limitations and computational challenges. In this study, we used Linear Inverse Modelling with an Ecopath model of the Irish Sea to investigate how the incorporation of uncertainty in dietary data can strengthen inferences based on model-derived ENA indicators. A Monte Carlo approach was used to generate ten thousand data-bound parameterisations for the Irish Sea food web and provide plausible distribution estimates for functional group diets. ENA results captured the plausible range of state-indicators and provided robust estimates of the control exerted by components within the food web. Results suggest that, higher trophic components, such as mammals, birds, and elasmobranchs in the Irish Sea are controlled by mid-to-low trophic components, such as small pelagic fish, invertebrates, and plankton. Fisheries discards also played an important role in the flow of energy to groups such as Nephrops (Norway lobster), crabs and lobsters, and seabirds. These results bolster our understanding of food web dynamics in the Irish Sea and demonstrate how information derived from ENA indicators can have implications for effective and sustainable ecosystem based management. Finally, the methods established here represent an important step in the maturation of marine ecosystem modelling and ENA for management purposes.