Modeling selenium bioaccumulation through arthropod food webs in San Francisco Bay, California, USA
Trophic transfer is the main process by which upper trophic level wildlife are exposed to selenium. Transfers through lower levels of a predator's food web thus can be instrumental in determining the threat of selenium in an ecosystem. Little is known about Se transfer through pelagic, zooplankton‐based food webs in San Francisco Bay ([SFB], CA, USA), which serve as an energy source for important predators such as striped bass. A dynamic multipathway bioaccumulation model was used to model Se transfer from phytoplankton to pelagic copepods to carnivorous mysids ( Neomysis mercedis ). Uptake rates of dissolved Se, depuration rates, and assimilation efficiencies (AE) for the model were determined for copepods and mysids in the laboratory. Small (73‐250 μm) and large (250‐500 μm) herbivorous zooplankton collected from SFB ( Oithona/Limnoithona and Acartia sp.) assimilated Se with similar efficiencies (41‐52%) from phytoplankton. Mysids assimilated 73% of Se from small herbivorous zooplankton; Se AE was significantly lower (61%) than larger herbivorous zooplankton. Selenium depuration rates were high for both zooplankton and mysids (12‐25% d −1 ), especially compared to bivalves (2‐3% d −1 ). The model predicted steady state Se concentrations in mysids similar to those observed in the field. The predicted concentration range (1.5‐5.4 μg g −1 ) was lower than concentrations of 4.5 to 24 μg g −1 observed in bivalves from the bay. Differences in efflux between mysids and bivalves were the best explanation for the differences in uptake. The results suggest that the risk of selenium toxicity to predators feeding on N. mercedis would be less than the risk to predators feeding on bivalves. Management of selenium contamination should include food webs analyses to focus on the most important exposure pathways identified for a given watershed.