Abstract
A two-compartment one-toxin model was fitted to data of the toxin level of Pacific oysters contaminated by Alexandrium tamarense in Hiroshima Bay, Japan. The model contained state variables for the viscera and non-viscera components of the oyster and was run using data for A. tamarense cell numbers with and without water temperature. It was tuned against toxicity data collected for the oyster. The fit of the predicted oyster toxicity to real data was best when there was no temperature link to cellular toxicity. Risk-assessment analysis suggests that the present model provides a robust indicator of the toxic status of the Pacific oyster in Hiroshima Bay. The model prediction of the toxin levels with routine monitoring of the toxic phytoplankton cell number could complement the monitoring of the toxin levels in shellfish by currently used time-consuming bioassays.
