Examination of ecological shifts using Bayesian hierarchical structural equation modeling

Session: Harmful Algal Blooms and Their Toxicity: Remote Sensing and Modeling Approaches (3)

Yuko Shimoda, University of Toronto Scarborough, [email protected]
Marten Koops, Fisheries & Oceans Canada, [email protected]
Monir Hossain, Fisheries and Oceans Canada, [email protected]
Shan Mugalingam, Lower Trent Conservation, [email protected]
George Arhonditsis, University of Toronto Scarborough, [email protected]

Abstract

The invasion of dreissenid mussels has been identified as one of the main culprits for the structural changes induced to the biotic communities and the persistence of eutrophication in the Bay of Quinte, Lake Ontario, after the mid-1990s. The colonization of mussels coincided with elevated nearshore TP levels, frequent cyanobacteria blooms, and distinct declining trends in species-specific fish biomass. In this study, we used structural equation modeling (SEM), parameterized with Bayesian hierarchical techniques, to draw inference regarding the likelihood of changes in the trophic relationships between pre- and post-dreissenid periods. Recent shifts in phytoplankton community composition were mainly associated with increased frequency of occurrence of toxin-producing Microcystis and reduced biovolume of N₂ fixers, such as Aphanizomenon and Anabaena. Our analysis elucidates the importance of different abiotic factors (light attenuation, water temperature, phosphorus, and ammonium) on total cyanobacteria, and Microcystis relative biovolume. We also found that fish biomass is shaped by the year-to-year TP variability, but this relationship has been modulated by various ecological events with the consequence that the Bay of Quinte fish assemblage has changed and the food web now produces less fish biomass per unit of TP.