MEPS prepress abstract  -  DOI: https://doi.org/10.3354/meps12504

A multi-model approach to understanding the role of Pacific sardine in the California Current food web

Isaac C Kaplan*, Tessa B Francis, André E Punt, Laura E Koehn, Enrique Curchitser, Felipe Hurtado-Ferro, Kelli F Johnson, Salvador E. Lluch. Cota, William J Sydeman, Timothy E Essington, Nathan Taylor, Kirstin Holsman, Alec D MacCall, Phillip S Levin

*Email: Isaac.Kaplan@noaa.gov

ABSTRACT: We develop a multi-model approach to explore how abundance of a forage fish (Pacific sardine Sardinops sagax) impacts the ecosystem and predators in the California Current, a region where sardine and anchovy Engraulis mordax have recently declined to less than 10% of contemporary peak abundances. We developed or improved applications of three ecosystem modeling approaches: Ecopath, MICE, and Atlantis. We also used Ecopath diets to predict impacts to predators using a statistical generalization of the dynamic Ecosim model (PREP). Models that included brown pelican Pelecanus occidentalis at the species level (MICE and Ecopath / PREP) both predict moderate to high vulnerability of brown pelicans to low sardine abundance. This vulnerability arises because sardine comprises a large fraction of their diet, and because other important prey (anchovy) also exhibit large population fluctuations. Two of the ecosystem models (MICE and Atlantis) suggest that California sea lions Zalophus californianus exhibit relatively minor responses to sardine depletion, due to having broader diets and lower reliance on another fluctuating species, anchovy. On the other hand, Ecopath/PREP suggests that sardine declines will have a stronger impact on California sea lions. This discrepancy may in part reflect structural differences in the models: Atlantis and MICE explicitly represent density dependence and age-structure which can mitigate effects of prey depletion in these models. Future work should identify fisheries management strategies that are robust to uncertainties within and among models, rather than relying on single models to assess ecosystem impacts of management and forage fish abundance.