Modelling the dietary exposure of SLE beluga to persistent organic pollutants

Jenny Oh, Leslie Saunders, James Armitage, Frank Wania

The elevated levels of persistent organic pollutants (POPs) in beluga whales (Delphinapterus leucas) of the St. Lawrence Estuary (SLE), classified as an endangered species, may be linked to their current population stagnation, despite many recovery efforts. However, a quantitative assessment of the bioaccumulation of POPs in the aquatic food web leading up to the SLE beluga whales has not yet been done. Therefore, further understanding the diet of the SLE belugas and modeling the bioaccumulation of contaminants through their food web may help shed light on the pathways of dietary exposure of these whales. Here, we construct a food web with feeding relationships and diet compositions of male and female SLE beluga whales using previously reported stomach contents data and staple isotopic ratio measurements (i.e., δ13C, δ15N) in beluga muscle tissue. Eleven potential prey species of the belugas were identified, whose feeding relationships, diet composition, and trophic position were estimated from the diets reported for each species in the literature. Using these feeding preference matrices of the aquatic food web and organism-specific characteristics collected from the literature, we parameterized a previously published food web model, which utilises a mass balance approach to calculate steady-state concentrations of lipophilic chemicals in each of the organisms. The final assembled food web model includes phytoplankton, zooplankton (copepods and krill), seven invertebrate species, nine fish species, and different age classes and sexes of beluga whales (adult and juvenile male, adult and juvenile female, infants). We applied the model to quantify the bioaccumulation of polychlorinated biphenyls and perfluoroalkyl sulfonamides in the food web of the SLE belugas. We also compare the predicted results with measured concentrations of these compounds in the whales and other aquatic species of the SLE to evaluate model performance.